Vol 9 (2017/18)

Patel and Ochroid: Comparison and Attempts at Reconciliation

By Nicholas Chiang


With its origins in contract law, Illegality has since been extended by the courts to other areas of law. Attempts to universalise the doctrine have led to considerable difficulty and inconsistency. The House of Lords decision of Patel v Mirza1 [Patel] was a valiant attempt to homogenise the approach to Illegality across all areas of law with a “range of factors” approach – but not without significant resistance, especially from jurists who prefer the deeply entrenched “rule-based” approach.2

Much closer to home, the Court of Appeal in the decision of Ochroid Trading Ltd v Chua Siok Lui3 [Ochroid] lays bare the difficulties in attempting to homogenise the doctrine, hinting that the ideal may indeed be a pipe dream. This article will examine the hurdles faced by the courts in formulating a universal approach in the law by comparing the English and local approach to illegality and discuss their compatibility.


The majority in Patel laid out a “range of factors” approach which applies to common law illegality. Rejecting a mechanistic approach based on the reliance rule, Lord Toulson held that the courts should consider a) the underlying purpose of the prohibition which has been transgressed, b) any other relevant public policies which may be rendered ineffective or less effective by denial of the claim, and c) the “possibility of overkill unless the law is applied with a sense of proportionality.”4

A more measured ”middle ground” was also adopted by Lord Neuberger, as the Singapore Court of Appeal observed in Ochroid. His Lordship held that there is a “rule” which generally entitles the plaintiff to restitutionary relief,5 and approved of the “range of factors” in determining when exceptions should be made to this “rule”. This departed from the majority’s approach to ascertain the validity of the contract on a discretionary “range of factors”. The distinction between 1) if the contract was void due to illegality and 2) if restitutionary relief should be granted was also made by the court in Ochroid.6 It is thus unclear what Lord Neuberger’s approach is to the former question.

The minority, however, preferred a rule-based approach. Lord Sumption, with whom Lord Clarke agreed, endorsed the reliance test. Under this approach, the validity of a claim turned solely on “whether the person making the claim is obliged to rely in support of it on an illegal act on his part”.7 His Lordship clarified that the reliance rule does not relate to procedural matters. Similarly, the court in Ochroid recognised two conceptions of the reliance rule: the first refers to reliance in the procedural sense, whereas the other refers to the principle in a normative or substantive sense when a claimant seeks enforcement of the illegal contract through the claim.8 The latter conception is good law in Singapore.9 It is noteworthy that the principle is only engaged when a claimant seeks to enforce an illegal contract, but is inapplicable when restitution is being sought.10

The minority dissent was premised on the fact that a “range of factors” approach grants the courts too much discretion, and removes the need for the application of legal rules.11 Furthermore, the majority approach engenders uncertainty due to the incommensurability of various factors and its applicability depends on the nature or degree of the illegality. Lastly, his Lordship pointed out the “range of factors” approach was unnecessary because the reliance rule is able to achieve substantial justice in the majority of cases. Arbitrariness in applying the reliance rule can be mitigated by the availability of restitutionary remedies.

The local approach since Ting Siew May v Boon Lay Choo12 [Ting Siew May] is a discretionary assessment of the proportionality of vitiating a claim in response to a pleaded illegality. It is an approach that balances a range of factors similar to the majority in Patel, albeit with slightly different factors.13 However, while the Patel “range of factors” approach extends to all forms of common law illegality, the discretionary approach in Ting Siew May is to be applied only in the “residuary common law category… of contracts entered into with the object of committing an illegal act”.14

The Patel “range of factors” approach was rejected in Ochroid for three reasons.

Firstly, the approach was unprincipled because it created an arbitrary distinction between common law and statutory illegality by granting unfettered discretion to the courts over illegality at common law.15 If a court is of the view that a contract should not be void pursuant to an established common law category of illegality, the court ought to “reconsider that particular category altogether”.16

Secondly, Phang JA also questioned the necessity of the approach in achieving justice as the availability of restitutionary remedies is sufficient to mitigate the harshness of the traditional rule. The Ting Siew May approach, with its proportionality principle, could also allow the court to reject the defence of illegality to avoid disproportionate and unjust results.17

Lastly, the uncertainty engendered by the “range of factors” approach was left unaddressed by the majority in Patel. These were, namely, that the balancing exercise involves uncertainty through the actual balancing exercise itself, and that the factors are incommensurate with no one factor being determinative. These problems according to Phang JA, were less severe under the approach in Ting Siew May because firstly, it was confined to “contracts entered with the object of committing an illegal act” and secondly, was limited by the “overarching principle of proportionality”.18 However, such uncertainty was unacceptable outside that restricted context.

While the court in Ochroid unequivocally rejected the “range of factors” approach, it is submitted that because of the overlapping similarity of approaches there is a semblance of universality – at least in the realm of “contracts entered into with the object of committing an illegal act”. However, this optimism is arguably short lived , given that parties to a void contract often seek restitutionary relief - wherein a different set of rules and principles apply.


With the majority’s “range of factors” approach Patel extending to civil claims of all sorts, the distinction between types of claims becomes irrelevant. The same approach would be applied to restitutionary claims.

While the minority in Patel was more circumspect with respect to contractual illegality, the opposite can be said with respect to restitution. For Lord Clarke, the harshness of the reliance rule in voiding illegal contracts could be ameliorated by the availability of restitutionary relief which, ought to be granted as long as restitutio in integrum can be achieved.19

On the doctrine of locus poenitentiae, Lord Mance construed it as a general principle of rescission that “puts parties back in the position that they should have been in…but for the entry into the contract which was or became affected and unenforceable by reason of the illegality”.20 His lordship also did not restrict the principle by reference to whether the illegal purpose has been carried out. Similarly, Lord Sumption stated that there is a general right to the restitution of money paid under an illegal contract. This position is similar to Lord Neuberger’s endorsement of the “Rule” granting a general right to restitution.

Dicta in Ochroid diverges from both the majority and minority in Patel.21 Phang JA rejected the minority view – more specifically, Lord Sumption’s position that restitutionary relief ought to be extended even to monies paid under a contract to commit murder. The learned Justice of Appeal characterised Lord Sumption’s view as one that would produce offensive or objectionable outcomes that would “bring the court into disrepute and undermine the integrity of the law”. The majority approach was also dismissed with the same reasons vis-a-vis contractual illegality.

In contrast to Lord Clarke’s general principle of rescission which makes available restitutionary relief as long as restitutio in integrum can be achieved, a claim in unjust enrichment in the local context would however be subject to the defence of illegality and considerations of public policy – which is distinct from the defence of illegality in a contractual sense.22 The court in Ochroid further adopted the principle of stultification which prevents restitutionary recovery if doing so would “undermine the fundamental policy that rendered the underlying contract void and unenforceable in the first place [emphasis added]”.23 This position is premised on the inadequacy of the approach which grants a general right to restitution. Even if a restitutionary claim does not provide the plaintiff with a remedy identical to contractual performance under the illegal contract, there are nevertheless remaining dangers of stultification; permitting recovery might provide future plaintiffs with a “lever” that the plaintiff might use to extract contractual performance from the defendant, or it might produce a “safety-net” for parties looking to engage in the same sort of illegality.

On the doctrine of locus poenitentiae, Phang JA gave two reasons for rejecting the Patel minority’s position. The minority had determined that the doctrine should be relaxed as a “general principle of rescission”, such that recovery was permitted as long as restitutio in integrum can be achieved. In response, Phang JA first observed that such a broad and relaxed doctrine could undermine the need for the need for the claimant to establish the requirements for a claim in unjust enrichment. It would, more importantly, allow the claimant to avoid the stultification principle.24

Secondly, Phang JA objected that a broad doctrine is incompatible with the role of the doctrine itself – namely to encourage timely withdrawal from illegality.25 Hence, Phang JA emphasised the need for genuine withdrawal, as opposed to mere voluntary withdrawal. In sum, the court characterised the doctrine as one which should only be invoked based on the policy of discouraging unlawful conduct.


While the “range of factors” approach has been extended to cover tortious claims in UK, the law is less settled in Singapore. The courts have traditionally used the close connection test, which according to Ooi Han Sun v Bee Hua Meng26 has a very limited application in tort – confining the test to situations where the injury is directly incurred in the course or commission of crime.27 The High Court in the recent decision of Md Shohel Md Khobir Uddin v Chen Yongbiao and another28 endorsed both the close connection test and the public policy approach of Hounga v Allen29 [Hounga] which can be considered the precursor of the test set out in Patel.

There are four possible paths that local courts may adopt to develop the illegality doctrine in tort law:

(1) Adopting the Patel “range of factors” approach.

(2) Continuing to use the close connection test.

(3) Attempting to reconcile both approaches.

(4) Extending the Ting Siew May approach to tortious claims.

Extending Ting Siew May’s approach (item (4), above) seems unlikely because the Ting Siew May approach is tailored for application to contractual illegality. For instance, one of the factors to be considered is the centrality of the illegality to the contract, which cannot be readily applied to tortious claims. It is suggested that an attempted reconciliation of the “range of factors” and “close connection” tests (item (3), above) will be the most viable way forward because the local courts are unlikely to only apply the Patel approach for the reasons set out above. However, the Patel approach (item (1), above) is potentially incompatible with the close connection test as it gives the courts the discretion to ignore the rule completely.

One plausible solution can be found in a flexible application of the close connection test, where policy considerations can influence the courts to either broaden or narrow the application of the test. In the case of Hounga, the Respondent (who was the employer) attempted to rely on the illegal employment of the Appellant (who was a victim of human trafficking) as a defence. With a flexible application of the close connection test, the Court could have applied the close connection test narrowly and find that the plaintiff did not fall within its scope and was thus entitled to recovery. The impetus behind a narrow application would naturally be overriding policy considerations, such as of discouraging human trafficking and exploitation in Hounga.

[1] [2016] UKSC 42.

[2] See e.g. Lords Sumption’s criticisms of a pliable “range of factors” approach to illegality in Les Laboratoires Servier & Anor v Apotex Inc & Ors (Rev 1) [2014] UKSC 55 at [19]-[21].

[3] [2018] 1 SLR 363; [2018] SGCA 5.

[4] Supra note 1 at [101].

[5] Ibid at [145].

[6] Supra note 2 at [64] – [65].

[7] Supra note 1 at [234].

[8] Supra note 2 at [128].

[9] As accepted in Ting Siew May v Boon Lay Choo and another [2014] 3 SLR 609; [2014] SGCA 28, at [127].

[10] Supra note 2 at [137].

[11] Supra note 1 at [261] – [264].

[12] Supra note 8.

[13] Ibid at [68].

[14] Supra note 2 at [110].

[15] Supra note 2 at [114].

[16] Ibid at [118].

[17] Ibid at [119].

[18] Ibid at [123].

[19] Supra note 1 at [220].

[20] Ibid at [193].

[21] Supra note 2 at [145].

[22] Ibid at [139] and [143].

[23] Ibid at [145].

[24] Ibid at [172].

[25] Ibid at [173].

[26] [1991] 1 SLR(R) 922; [1991] SGHC 73.

[27] Ibid at [15]

[28] [2018] 3 SLR 160; [2017] SGHC 109.

[29] [2014] UKSC 47; see supra note 25 at [33]-[34].

The PDF version of this article is available for download here.

The Trojan Horse of Therapeutic Privilege in Hii Chii Kok

By Naomi Koh Jie Min


The 2017 Court of Appeal case Hii Chii Kok v Ooi Peng Jin London Lucien1 [Hii Chii Kok] has been heralded as “[t]he end of the ‘doctors know best’ era” by commentators.2 In a landmark decision, a Coram of five Judges unanimously rejected the unapologetically paternalistic Bolam-Bolitho3 test for negligent advice in favour of a three-stage test based on the UK Supreme Court’s 2015 decision in Montgomery v Lanarkshire Health Board4 [Montgomery]. As the Court put it, a “seismic shift in medical ethics, and in societal attitudes towards the practice of medicine” necessitated an approach that balances beneficence with patient autonomy.5 Hii Chii Kok is indisputably a welcome addition to local case law on medical negligence, but it should not be welcomed uncritically. A key issue with Hii Chii Kok lies in the unclear scope of the defence of therapeutic privilege6 to a claim of negligent advice. In fairness to the Court of Appeal, the doctrine of therapeutic privilege, as Mulheron observes, “has suffered from an almost complete lack of judicial delineation in English law”7 and by extension, Singapore law. However, the cost of leaving therapeutic privilege undefined is high; it leaves the Hii Chii Kok test open to abuse.


A. Singapore: Gunapathy

To understand Hii Chii Kok, one must consider the state of the law that preceded it. Locally, Khoo James v Gunapathy d/o Muniandy8 [Gunapathy] stood for the position that a two-stage inquiry was required to determine whether a professional had breached his duty of care: namely, whether the expert (1) directed his/her mind to the comparative risks and benefits, and (2) reached a “defensible conclusion”9 after balancing risks and benefits. In other words, the Court of Appeal had adopted the Bolam test of the ordinary skilled man exercising and professing to have a skill,10 with the Bolitho addendum that the conclusion reached must have a logical basis.11 Further, and more pertinently, the Court accepted that the Bolam-Bolitho test applied to pre-treatment advice, following the House of Lords in Sidaway v Bethlem Royal Hospital12 [Sidaway].

B. UK: Sidaway Overuled in Montgomery

Montgomery amended the standard for risk disclosure to one where the doctor is:

“[U]nder a duty to take reasonable care to ensure that the patient is aware of any material risks involved in any recommended treatment, and of any reasonable alternative or variant treatments. The test of materiality is whether, in the circumstances of the particular case, a reasonable person in the patient’s position would be likely to attach significance to the risk, or the doctor is or should reasonably be aware that the particular patient would be likely to attach significance to it.”13

In this decision, the House of Lords effectively overruled the test in Sidaway. Although, as Hobson observes, their Lordships tried to “frame Montgomery as a development of Sidaway ... to show that the development from Sidaway to Montgomery is smooth”,14 in principle Montgomery effectively displaces the concerns and policies that undergirded Sidaway’s stance that Bolam-Bolitho applied to medical advice.


In Hii Chii Kok, the Court acknowledged the increasing recognition of patient autonomy across common law jurisdictions.15 More specifically, the scope of patient autonomy is greater in relation to medical advice, which guides patients to make an active choice.16 Hence, the Bolam-Bolitho test was no longer applicable in relation to advice. Rather, the Court endorsed a three-stage test (referred to as the “Hii Chii Kok test”).

In the first stage of the Hii Chii Kok test, the Court answers the question of what information the doctor has a duty to disclose. This comprises of what a reasonable person in the patient’s position would wish to know, as well as information that the doctor knows or ought to have known would be important to the particular patient in question.17 The first stage therefore effectively echoes Montgomery.

At the second stage, the Court evaluates whether the doctor was in possession of the pertinent information. If answered in the negative, the Court then has to assess whether the doctor’s ignorance arose from negligence in diagnosis and treatment. This inquiry would fall under the ambit of the Bolam-Bolitho18 test.

If the second stage is answered in the affirmative, then the final question is whether the doctor’s withholding of information was justified. The Court declined to limit the scope of potential justifications, but identified emergencies, waivers, and therapeutic privilege as specific instances in which a doctor may withhold information from a patient.19 It is the exception for therapeutic privilege that is contentious as further explored below.


A. The Exception

Although Hii Chii Kok purports to eschew the Bolam-Bolitho test in favour of a more expansive paradigm with greater room for patient autonomy, questions have been raised as to the efficacy of the revised test in promulgating these goals. While the Hii Chii Kok test is a step forward with a more patient-focused approach, the wide yet murky scope of the therapeutic privilege exception creates a very real concern. If not properly defined, this exception leaves an opening for medical paternalism to reassert itself. As discussed below, doctors could infringe on a patient’s right to make an informed decision under the guise of the patient’s supposed inability to do so.

What do we know about the therapeutic privilege exception? The Court of Appeal has suggested that it applies exclusively in “exceptional circumstances”,20 but has been vague on the precise details of the test. Ostensibly, it is an objective factual inquiry21 as to whether the patient is likely to be harmed disproportionally to the impact of being provided with the information in question. Showing advertence to the reality that the therapeutic privilege exception is prone to abuse, the Court explicitly stated that the exception should not preclude patients from making a choice merely because the patient’s decision appears contrary to their interests.22 However, and albeit discrepantly, patients who may choose not to undergo “relatively safe treatments that [could] drastically improve their quality of life” may be caught under the exception.23 Moreover, patients who possess mental capacity but also have impaired decision-making abilities are caught under this exception.24

The Court’s attempt to delineate the boundaries of therapeutic privilege is unsatisfying. For one, must the “serious physical or mental harm”25 be medically recognised? Further, would “physical or mental harm” capture scenarios where the anticipated harm arises from the patient’s anticipated refusal to undergo treatment? Given the Court of Appeal’s observation that therapeutic privilege would include “certain geriatric patients who … may be ‘easily frightened out of having even relatively safe treatments that can drastically improve their quality of life’”,26 this appears to be the case. However, this statement is problematic, especially in light of the potentially conflicting Mental Capacity Act27 [MCA] as discussed below.

The confusion that the therapeutic privilege exception generates is exacerbated by the reality that the three stages of diagnosis, advice, and treatment are not mutually exclusive. As the Court itself in Hii Chii Kok noted, “a single step in the medical care process will engage more than one aspect of the doctor’s duty, and the different aspects will then be in play concurrently”28. This permeability between stages is an issue given that it affects the test to be applied and consequently whether the therapeutic privilege exception would even come in play.

This problem was made apparent in Noor Azlin Bte Abdul Rahman v Changi General Hospital Pte Ltd29 [Noor Azlin]. In that case, one of the issues in dispute was whether the alleged negligence fell under the scope of diagnosis and treatment, or advice. The plaintiff argued that the defendants failed to advise her of the risk of cancer, such that she was unaware that other diagnostic options existed and did not make any further investigations. Nonetheless, the High Court found for the defendant. The lack of such advice stemmed from the defendants’ professional diagnosis that cancer was not a likely cause of the plaintiff’s complaints. Hence, the proper test to be applied was the Bolam-Bolitho test. Although the decision in Noor Azlin is not in itself controversial, the lack of a clear directive on how the stages are to be distinguished is troubling.

B. Therapeutic Privilege and the Mental Capacity Act

The MCA was enacted to regulate the decision-making process on behalf of those who lack capacity and hence are unable to make decisions for themselves. At first blush, the Hii Chii Kok test has no ostensible relation to the MCA; the former addresses the question of whether the patient’s consent was informed, while the latter deals with treatment in a situation where the patient is incapable of consenting.

However, it is unclear whether a doctor may evade liability for non-disclosure under the MCA in specific situations. More specifically, in light of the undefined boundaries of the therapeutic privilege exception, the implications of the MCA on the Hii Chii Kok test should be explored given that both fundamentally affect whether the principle of patient autonomy can be overridden by the patient’s lack of decision-making capacity. Under the Hii Chii Kok test, therapeutic privilege is explicitly raised as a defence to a charge of negligent advice where “beneficence comes to the fore”,30 echoing the MCA requirement that the decision be made in the “best interests”31 of the incapacitated individual.

Under s 7(2) of the MCA,32 a defendant doctor would not incur liability for an act done in connection with the care and treatment of the plaintiff, assuming that his or her act would not have attracted liability if the plaintiff hypothetically could and did consent. The defendant medical professional must take reasonable steps to establish whether the plaintiff lacks capacity. In addition, he or she must have reasonably believed that the plaintiff lacks capacity, and that it is in the plaintiff’s best interests that the act be done.33 It is conceivable that the “act” in this situation would extend to non-disclosure on the basis of therapeutic privilege. Which acts, therefore, would fall under the ambit of the MCA, and which would fall under the ambit of the Hii Chii Kok test?

The distinction between the two appears to be one of degree. For the therapeutic privilege exception, the Court suggested that it “should extend to cases where although patients have mental capacity, their decision-making capabilities are impaired to an appreciable degree34 [emphasis added]. In contrast, the MCA arguably requires total incomprehension35 or total inability36 to use or weigh information in the process of decision-making. Ultimately, however, the outcome regardless of which is applied is the same: both act to relieve the defendant of tortious liability.

Crucially, however, the MCA explicitly provides that a person is not to be treated as unable to make a decision merely because said decision is unwise.37 This point has been made earlier by Low, who argues that an “informed decision is not synonymous with [a] good decision”.38 This stands in stark contrast to the principle of patient autonomy apparently espoused by Hii Chii Kok – namely, that a doctor may justify omissions on the basis that he or she believed that the patient would be “frightened out of having even relatively safe treatments that can drastically improve their quality of life”.39

Clearly, a successful finding that the patient was fully incapacitated under the MCA is significantly more onerous than the finding that the patient was of a class to trigger the therapeutic privilege exception under Hii Chii Kok. Consequently, it should stand to reason that the corresponding boundaries of the test in Hii Chii Kok should be narrower. Since both exculpate a doctor facing a charge of negligence, it would be logically inconsistent to allow a doctor to prove less under the Hii Chii Kok test.


Ultimately, the therapeutic privilege exception as it stands is problematic because it is unclear. While it is meant to relate to “exceptional”40 situations, Hii Chii Kok does not provide a workable litmus test for when a situation would fall within its boundaries. Rather, Hii Chii Kok appears to take an expansive view of therapeutic privilege – perhaps more so than other jurisdictions. Pertinently, the House of Lords in Montgomery expressly rejected41 the defendant doctor’s argument in that case that it was “not in the maternal interests for women to have caesarean sections”42.

Granted, it is within the bounds of plausibility that a patient could be clearly and utterly incapable of making the required decision, but not trigger the provisions of the MCA. There reasonably is a space for the therapeutic privilege exception – but the current boundaries of the test are too wide. With respect, further clarification is essential.

In fairness to the Singapore Court of Appeal, the problems surrounding the doctrine of therapeutic privilege apply across jurisdictions. As Mulheron establishes, “the reality is that there is one clear instance of therapeutic privilege in English law”, and this rarity of application hampers the creation of a coherent, articulate doctrine.43

However, this rarity does not justify the continuing confusion surrounding therapeutic privilege. Locally, about 10% of the population suffers from anxiety and depressive disorders.44 This segment – along with the overly anxious, the hesitant to undergo medical procedures, and perhaps even the simply cautious – risk being caught unnecessarily by too expansive a doctrine. Ultimately, for Hii Chii Kok to truly represent a shift in medical ethics towards upholding the principle of patient autonomy, the therapeutic privilege doctrine must be re-considered.

[1] [2017] SGCA 38; [2017] 2 SLR 492.

[2] Hairul Hakkim and Kevin Ho Hin Tat, “The end of the ‘doctors know best’ era – from medical paternalism to patient autonomy”, Singapore Law Blog, (16 July 2017) online: <http://www.singaporelawblog.sg/blog/article/191> accessed 10 June 2018.

[3] Supra note 1.

[4] [2015] UKSC 11; [2015] AC 1430.

[5] Ibid at [120].

[6] Also termed “doctor’s privilege” or “therapeutic exception” in some cases.

[7] Rachael Mulheron, “Has Montgomery Administered the Last Rites to Therapeutic Privilege? A Diagnosis and a Prognosis” (2017) 70:1 Current Leg Probs at 149.

[8] [2002] SGCA 25; [2002] 1 SLR(R) 1024.

[9] Defensible is not to be equated with reasonable. See Gunapathy, ibid at [65].

[10] Specifically, “[a doctor] is not guilty of negligence if he has acted in accordance with a practice accepted as proper by a responsible body of medical men skilled in that particular art” [emphasis added]: Bolam v Friern Hospital Management Committee [1957] 1 WLR 582 at 587.

[11] Yeo Peng Hock Henry v Pai Lily [2001] SGCA 72; [2001] 3 SLR(R) 555 at [16], as affirmed in Gunapathy at [52], [61].

[12] [1985] AC 871 (HL).

[13] Supra note 4 at [87].

[14] Clark Hobson, “No (,) More Bolam Please: *Montgomery v Lanarkshire Health Board*” (2016) 79(3) MLR 468.

[15] Supra note 1 at [116].

[16] Ibid at [93].

[17] Ibid at [137], [138].

[18] Ibid at [133], [147].

[19] Ibid at [134], [149] and [151].

[20] Supra note 1 at [153].

[21] Ibid.

[22] Ibid.

[23] Ibid at [152].

[24] Ibid at [153].

[25] Ibid at [152].

[26] Ibid.

[27] Cap 177A, 2010 Rev Ed Sing.

[28] Supra note 1 at [90].

[29] [2018] SGHC 35.

[30] Supra note 1 at [149].

[31] Supra note 28, s 6(10).

[32] Ibid, s 7(2).

[33] Ibid, s 7(1).

[34] Supra note 1 at [152].

[35] Supra note 28, s 5(1)(a).

[36] Ibid, s 5(1)(c).

[37] Ibid, s 3(4).

[38] Kee Yang Low, “Doctor’s duty of disclosure and the Singapore Court of Appeal decision in Hii Chii Kok: Montgomery transformed” [2017] 25:2 Tort L Rev 79 at 89.

[39] Supra note 1 at [152].

[40] Ibid at [153].

[41] Supra note 4 at [91].

[42] Ibid at [13].

[43] Supra note 7, at 186.

[44] Jaclyn Lim, “Anxiety in Singapore: Stats, Types and Who’s at Risk”, HealthXchange.sg (blog), online: <https://www.healthxchange.sg/wellness/mental-health/anxiety-singapore-stats-types-risk> accessed 12 July 2018.

Enforcement of Arbitral Awards Set Aside at the Seat of Arbitration: The Way Forward for Art. V(1)(e) in Singapore

By Daniel Ang Wei En*


A. Singapore's Pro-Enforcement Policy

The Singaporean courts interpret the statutory grounds for setting aside awards narrowly1 and with strict scrutiny.2 This approach is consistent and seeks to protect the sanctity of the arbitral award.3 In TMM Division Maritima SA de CV v Pacific Richfield Marine Pte Ltd,4 Chan Seng Onn J summed up the Singapore position that “the power … to set aside awards, must and should only be exercised charily”. Singapore’s approach represents the mainstream curial philosophy across the globe.5

B. Extent of Singapore’s Pro-Enforcement Judicial Attitude

For the Singapore courts, dealing with a seat court’s setting aside of the award raises the question of whether to focus on: (a) the award itself; as opposed to (b) the process and effects of the seat court’s decision to set aside.

French courts have demonstrated a clear deference to the arbitral award, which is perceived as self-sufficient, constrained only by French law.6 On the other hand, the US courts have refused enforcement based on why the award was set aside.7

Given Singapore’s standing as a global-leading arbitration centre, the approach that the Singapore courts take is crucial in contributing to the international jurisprudence on this matter, for which there is no clear litmus. Singapore’s position could possibly nudge the divided international community towards a more uniform enforcement of arbitral awards.


The relevant provision is Article V(1)(e) of the Convention on the Recognition and Enforcement of Foreign Arbitral Awards (New York, 1958)8 [New York Convention], which formed Article 36(1)(a)(v) of the UNICTRAL Model Law on International Commercial Arbitration9 [Model Law]. The latter was incorporated in Singapore’s legislation through Section 31(2)(f) of the International Arbitration Act10 [IAA]. Therefore, these provisions should be interpreted harmoniously.


A. Grounds for Refusal in the International Arbitration Act

The IAA sets out three (3) grounds for refusal within Section 31(2)(f), where the award:

(a) has not yet become binding;

(b) has been suspended; or

(c) has been set aside.

At present, Section 31(2)(f) has not been pleaded as a ground for non-enforcement.11

B. Grounds for Refusal in Case Law

The Singapore Court of Appeal expressed “tentative thoughts” on this issue in PT First Media TBK v Astro Nusantara International BV 12 [PT First Media], which are obiter.13

(1) Rejection of the french authorities

Reviewing the French authorities of Hilmarton Ltd v Omnium de traitement et de valorisation14 [Hilmarton] and The Arab Republic of Egypt v Chromalloy Aeroservices, Inc 15 [Chromalloy], the Court of Appeal rejected the wider notion of “double control” as adopted by the French courts which17:(a) recognised that awards do not derive their validity from a particular local system of law, and (b) applied French legislation18 which did not contain the equivalent of Article V(1)(e) of the New York Convention.

(2) Implication and purpose of enforcement

In principle, the Court of Appeal in PT First Media “seriously doubted”19 enforcing an award set aside at the seat of arbitration on three grounds:(a) since the award derives legal effect from the law of the arbitral seat, its annulment means there is no award to enforce to begin with (the “legal order ground”), (b) Singapore’s domestic law does not confer “more favourable right[s]”20 of enforcement than Section 31(2)(f) of the IAA, (c) a purposive interpretation of the Article V(1)(e) of the New York Convention requires that the provision must have legal effect outside of the seat court (the “purposive interpretation ground”).

In addition to rejecting the French approach, the Court of Appeal appears to have adopted the US approach itself by considering enforcement based on the effects of the setting aside decision. Ground (b) is non-contentious. Grounds (a) and (c) will be explored below.


A. Purposive Interpretation of Article V(1)(e) of the Convention

Ascertaining the provision’s purpose is crucial, given the Article 31(1) of the Vienna Convention21 requirement of purposive interpretation.

The New York Convention sought in general to make the enforcement of arbitral awards easier22 and internationally uniform23, as mentioned by the English High Court in Dowans Holding S.A. v Tanzania Electric Supply Co. Ltd.24 [Dowans Holdings]. The particular purpose of Article V(1)(e) was to make enforcement less burdensome by removing the requirement of “double exequatur”.25 As explained by the Chairman of the Working Party, “it would be unrealistic to delay the enforcement of an award until all the time limits provided … had expired or all possible means of recourse … have been exhausted and the award had become “final”.”26

Hence, while the prior Geneva Convention27 required an award to be “final”, a deliberate choice was made to use the word “binding” in Article V(1)(e) instead. The drafters’ particular intent to eliminate “double exequatur” behind Art V(1)(e) has received judicial acceptance, not only in the English High Court's decision in Dowans Holdings28, but also by the Swiss Federal Tribunal in Y v X.29 It follows that enforcement cannot be automatically defeated merely because the seat court refuses to enforce the award.

B. Purpose of Article V(1)(e) in Relation to the Discretion of the Enforcement Court

Sundaresh Menon CJ opined in PT First Media30 that “[i]f [setting aside the award] would only ever be of efficacy in relation to enforcement proceedings in the seat court, then it seems to have been devised for little, if any, discernible purpose [emphasis added].” What then, is the implication of the setting aside of an award?

(1) Basis for the enforcement court’s discretion to enforce the award

Menon CJ’s opinion appears to proceed on the assumption that in the enforcement court, the award must be enforced even if the Article V(1)(e) ground is present, rendering the setting aside ineffectual outside the seat court. However, Article V(1)(e) uses the word “may” and thus confers discretion to enforce. The setting aside by the seat court does not automatically defeat the enforcement of the award elsewhere.

This poses a tension with the fact that the seat court’s setting-aside cannot be strictly ignored. The drafters of Article V(1)(e) did not intend for the setting-aside by the seat court to be completely immaterial to the enforcement court. This tension was explored by the Convention delegates:

“[Article V(1)(e)] reflects the inability of the Conference to agree on the solution to the problem of the “double exequatur”. No one wanted the Convention to require judicial proceedings in confirmation of the award in both the rendering and enforcing State. At the same time, an award which had been set aside … should hardly be granted enforcement in another State.” 31

It is therefore unlikely that Article V(1)(e) was intended to be ignored in favour of compulsory enforcement by the enforcement courts. While the delegates proposed limiting the seat courts' control,32 these limits were unspecified.33

(2) Exercising the discretion accorded to enforcement courts

Taking the US approach of scrutinizing the basis for the seat court’s setting aside, how much discretion does the enforcement court have resulting from its “tug-of-war” with the seat court, and how should it be applied?

(a) Awards set aside on local standards of annulment.

The delegates considered eliminating local standards for annulment, while retaining international standards for annulment such as the New York Convention’s grounds. That was intended to concentrate power in the hands of the enforcement courts, imposing a more uniform regime for the annulment of awards.34

Proponents of the “territorialist” approach have argued conversely that since the drafters wanted the seat court to conduct the principal review of the award,35 the enforcement court should defer to the seat court’s judgment instead. However, adopting this “territorialist” approach undermines arbitration as an effective international dispute resolution mechanism. It would require giving international legitimacy to local standards by importing them into the Convention, notwithstanding that local standards for setting aside awards may be inconsistent with the internationally-accepted standards36 – the sole standards the Convention purports to enforce. Consequently, recognising local standards would then result in internationally inconsistent application, as well as inconsistency in the law of the Convention itself. With respect, it is clear that asserting a “territorialist” position would be inconsistent with the Convention’s purpose of effecting greater uniformity of enforcement internationally.37

Additionally, local standards themselves could be perceived as improper or objectionable by the international community. Jan Paulsson illustrates this with a hypothetical example where an award is annulled because it violates a local rule that all members of the tribunal be men or of a particular religious confession.38

Therefore, enforcement courts should not refuse enforcement only because the award has been annulled according to local standards.39 Respecting the sovereignty of the seat should entitle the seat court to adopt local rules regarding the set-aside of awards to comply with local preferences, without necessarily having international effect.40

(b) Awards set aside on international standards of annulment

Conversely, courts have conversely refused to enforce awards because they were annulled according to international standards.

In the United States, the Federal District Court for the District of Columbia enforced the award in Re Chromalloy Aeroservices Inc. v The Arab Republic of Egypt41 [Chromalloy (US)] because the annulment grounds were domestic in that the award was “not properly grounded under Egyptian law”.42 However, the US Court of Appeals for the Second Circuit refused enforcement in Baker Marine (Nig.) Limited v. Chevron (Nig.) Limited, Chevron Corp., Inc and others v. Danos and Curole Marine Contractors, Inc.43 [Baker Marine] because the award was annulled on the international grounds in Article V(1)(c) and (d).44

The court in Baker Marine stated that enforcement according to domestic law “would seriously undermine finality and regularly produce conflicting judgments.”45 It is submitted that this should also apply to seat courts’ refusing enforcement according to their own domestic law.

For reasons of practicality, a compromise should be made between enforcing awards annulled on local grounds and refusing enforcement if the award was annulled on international grounds.


Menon CJ opined in PT First Media46 that “the contemplated erga omnes effect of a successful application to set aside an award would generally lead to the conclusion that there is simply no award to enforce”.

A. The Territorial Approach

This is the starting point of the “territorial” approach.47 Van den Berg argues that “[t]he fact that the award has been annulled implies that the award was legally rooted in the arbitration law of the country of origin.”48 Professor Pieter Sanders, a key drafter, appears to have agreed with the result.49

B. Issues with the Territorial Approach

(1) Problems with the territorial approach in general

The first objection is that the territorial approach is in direct opposition to the text of Article V(1)(e) of the Convention, which confers discretionary power to enforce the award.50

In addition, the territorial approach is grounded on an outdated assumption that the law of the seat court provides the award’s legal force.51 In 1958, when the Convention was drafted, the role of the arbitral seat was arguably more substantial that it is today.52 The seat of arbitration is agreed upon by parties mainly out of convenience or compromise, instead of the inherent importance of the arbitral seat itself.

Further, the parties’ agreement did not include submitting to the exclusive jurisdiction of a particular court. By agreeing to extra-curial arbitration, they arguably intended to avoid the reach of the seat court in the first place.53 The territorial approach contradicts the parties’ intentions as it allows seat courts to interfere with the legal effect of the award. It is doubtful if seat courts should arrogate to themselves powers to do this, and the territorial approach may assume too great a significance on the part of the seat court here.

(2) The Article VII(1) exception

The enforcement courts are empowered relative to the seat courts by Article VII(1) of the Convention, which allows for ‘opting' out of the Convention and into the relatively more pro-enforcement domestic law provisions of the enforcement court. Since Article VII(1) empowers the enforcement court to determine, unilaterally, whether the award would be enforced, the assumption that the award truly derives its legal effect solely from the legal order of the seat court is invalid.

In Hilmarton,54 the Paris Cour d’appel held that Article VII(1) prevailed over Article V, and thus applied55 Article 1502 of the New Code of Civil Procedure to enforce the award.

While Article VII(1) might be interpreted in an uncertain manner,56 it is undergirded by the interests of sovereignty which must prevail as a recognised pillar of the New York Convention. Uncertainty is thus a necessary but surmountable procedural cost. For example, the enforcement court is not bound by foreign judgments if they are not contrary to domestic public policy.57 More importantly, sovereignty is an inalienable principle of international law, while consistency is merely a desirable outcome. Uncertainty should not be inflated as an obstacle to the use of Article VII(1).

Inconsistency is less of a problem, given that enforcement courts are less deferential to foreign judgments that set aside the award than foreign judgments that decide the merits of the underlying dispute.58 Just as the US Supreme Court recognised the principle of international res judicata in Hilton v Guyot,59 the enforcement courts should defer to the authority that decides the merits of the case instead to avoid re-litigating the dispute elsewhere.60

(3) Burden of having endless enforcement proceedings all over the world

The application of Article VII(1), as noted in Baker Marine,61 has the result that “a losing party will have every reason to pursue its adversary ‘with enforcement actions from country to country until a court is found, if any, which grants the enforcement.’”

The practical extent of such enforcement actions is, however, limited. The award creditor will only seek enforcement in the countries containing assets of the award debtor.62 In fact, in Yukos Capital S.A.R.L v. OAO Tomskneft VNK,63 the lack of assets was undeniably relevant even in the early stages of enforcement. In (1) X1, (2) X2 v (1) Y1, (2) Y2,64 the Dubai International Financial Courts held that the lack of assets may be a ground to refuse the enforcement.

Further, the judgment creditor is likely to narrow the scope of his enforcement actions to countries where the judgment debtors’ assets are easier to be enforced against. For instance, it is particularly difficult to satisfy the judgement debt with illiquid assets.

C. The Delocalised Approach

The alternative to the territorial approach does lend support to the idea of focusing the inquiry on the award itself. The award was considered by the French Court of Cassation in PT Putrabali Adyamulia (Indonesia) v Rena Holding65 to be “not attached to any state legal order”, but a “decision of international justice whose regularity is examined according to the rules applicable in the country where its recognition is sought". This view ought to be followed, because parties who agree to international arbitration in Convention States assume that it is the Convention that serves as the basis for enforcement, and not another country’s domestic law.66

This is buttressed Professor Gaillard’s view that arbitrators “do not derive their powers from the State in which they have their seat, but rather from the sum of all the legal orders that recognise … the validity of the arbitration agreement and the award. …” Arguably, the “sum of legal orders” is embodied by the signatories' ratification of the Convention itself.


A. Implications in General

While the territorial approach has its difficulties, the delocalised approach finds little support internationally. Given that a compromise can be reached between the difficulties arising from scrutinizing the seat court’s decision to set aside, there is insufficient reason in principle for Singapore to contradict the weight of comity and uniformity by departing from the US approach. This is buttressed by the purposive interpretation of Article V(1)(e).

While sovereignty is the main feature of the delocalised approach, sovereignty is indeed respected even in the scrutiny of the seat court’s decision to set aside, without the same severe expense of certainty and uniformity in enforcement.

B. Implications for Singapore

As one of the world’s top international arbitration centres, the Singapore courts’ decisions inexorably drive the development of international arbitration jurisprudence. In this particularly divided area, with the French, US and other courts taking seemingly irreconcilable positions, Singapore’s answer is particularly pertinent in advancing international arbitration as a desirable mechanism of international dispute resolution.

The course charted by the Singapore Court of Appeal appears consistent with the goal of attaining uniformity through comity, but it remains to be seen how the Singapore courts will apply the approach. Should the Singapore courts adopt an “internationalist” stance on this matter, it would be highly persuasive in moulding the global enforcement regime to become more modern and cohesive.

* LL.B. (Hons) candidate, National University of Singapore. I record a debt of gratitude to T.G. Khoo, LL.B. (Coll. Reg. Lond.), LL.M. Candidate (Cantab.), whose invaluable learning has greatly benefited an earlier draft of this paper; and to Samuel Ang Rong En, LL.B. (Hons) candidate, NUS, for his keen insight and invaluable guidance. Any errors and infelicities are, necessarily, my own.

[1] See, “Trends of the 2017 Singapore review” (16 March 2017), Global Arbitration Review Know-How: Commercial Arbitration (blog), online: <http://globalarbitrationreview.com/know-how/results?question_ids=1000341&jurisdiction_ids=1004606&edition_id=1000012> (accessed 13 December 2017).

[2] Paul Tan & Alessa Pang, “The International Arbitration Review – Edition 8 (Singapore)” (August 2017), The Law Reviews (blog), online: <https://thelawreviews.co.uk/edition/the-international-arbitration-review-edition-8/1145747/singapore>.

[3] See Michael Hwang SC & Su Zihua, “Egregious Errors and Public Policy: Are the Singapore Courts too arbitration friendly?” (delivered at the Singapore Academy of Law Conference 2011 – Singapore Law Developments (2006–2010)) at pp. 19–55. For examples of Singapore’s firm pro-arbitration approach, see also David Williams QC, “Defining the Role of the Court in Modern International Commercial Arbitration” (Herbert Smith Freehills SMU Asian Arbitration Lecture, delivered at the Singapore Management University in 2012), online: < http://www.arbitrationconference.com/download/file/218/>.

[4] [2013] 4 SLR 972 at [1].

[5] See Sundaresh Menon SC (as His Honour then was), “International Arbitration - The Coming of a New Age for Asia (and Elsewhere)” (Conference paper delivered at the ICCA Congress 2012, Opening Plenary Session) at [5].

[6] See François-Xavier Train, “Interpretation and Application of the New York Convention in France”, in George Bermann ed., Recognition and Enforcement of Foreign Arbitral Awards: The Interpretation and Application of the New York Convention by National Courts (Springer, 2017) 281 at 304.

[7] 191 F (3d) 194 at 197  (2nd Cir 1999).

[8] 10 June 1958, 330 UNTS 38 (entered into force 7 June 1959).

[9] 24 June 2002, 24 ILM 1302 (1985).

[10] Cap 143A, 2002 Rev Ed Sing.

[11] Jean Ho, “The Interpretation and Application of the New York Convention in Singapore”, in George Bermann ed., Recognition and Enforcement of Foreign Arbitral Awards: The Interpretation and Application of the New York Convention by National Courts (Springer, 2017) 813, at 828.

[12] [2013] 1 SLR 372 at [76].

[13] Ibid at [77].

[14] (1995) XX Yearbook Comm Arb 663–665.

[15] (1997) XXII Yearbook Comm Arb 691–695.

[16] Supra note 11, at [76].

[17] Ibid at [77].

[18] Article 1502 of the French New Code of Civil Procedure, as referred to in the judgment.

[19] Supra note 11.

[20] See supra note 9, Article VII of the New York Convention.

[21] Vienna Convention on the Law of Treaties, 23 May 1969, 1155 UNTS 331 (entered into force 27 January 1980).

[22] Nadia Darwazeh, “Article V(1)(e)”, in Kronke et al., eds, Recognition and Enforcement of Foreign Arbitral Awards: A Global Commentary on the New York Convention (Kluwer Law International, 2010) 301, at 333.

[23] Ibid at 332.

[24] [2011] EWHC 1957 (Comm).

[25] Supra note 22, at 306.

[26] “United Nations Conference on International Commercial Arbitration, Summary Record of the Seventeenth Meeting,” UNESCOR, 1958, UN DOC E/CONF.26SR.17.

[27] Convention on the Execution of Foreign Arbitral Awards, 26 September 1927, 92 UNTS 301 (entered into force 25 July 1929).

[28] Supra note 24.

[29] Swiss Federal Tribunal, Switzerland, 3 January 2006, 5P.292/2005.

[30] Supra note 12, at [77].

[31] Leonard V. Quigley, “Accession by the United States to the United Nations Convention on the Recognition and Enforcement of Foreign Arbitral Awards” (1961) 70:7 Yale LJ 1069.

[32] ECOSOC, 1958, 11th Mtg, UN Doc E/Conf.26/SR.11 at 6.

[33] Emmanuel Gaillard & John Savage, eds, Fouchard Gaillard Goldman on International Commercial Arbitration (The Hague: Kluwer Law International, 1999) at 979.

[34] ECOSOC, Comments on Draft Convention on The Recognition and Enforcement of Foreign Arbitral Awards – Note of the Secretary-General, UN Doc E/Conf.26/2, March 1958) at 16-19.

[35] Supra note 22, at 327.

[36] Marike Paulsson, “Chapter 6: Resisting Enforcement of Awards”, in The 1958 New York Convention in Action (United States: Wolters  Kluwer, 2016) 157-216, at 212.

[37] Supra note 22, at 332.

[38] Jan Paulsson, “Enforcing Arbitral Awards Notwithstanding a Local Standard Annulment (LSA)” (1998) 9:1 ICC Intl Ct Arb Bull 14 at 17.

[39] Ibid at 25, 29.

[40] Ibid at 22.

[41] 939 F Supp 907 at 911 (D.D.C. 1996).

[42] Ibid.

[43] Supra note 7.

[44] This distinction is buttressed by the refusal to enforce in TermoRio S.A.E.S.P. (Colombia) and LeaseCo Group, LLC. V Electranta S.P (Columbia) (District of Columbia 2007), in Yearbook Commercial Arbitration XXXIII (2008) (United States no. 621), at 955-969, where the award was annulled under Article V(1)(a) and (2) of the Convention.

[45] Supra note 7, at 197.

[46] Supra note 14, at [77].

[47] Supra note 22, at 326.

[48] Van Albert Jan van den Berg, “Annulment of Awards in International Arbitration” in Richard B. Lillich & Charles N. Brower, eds, International Arbitration in the 21st Century: Towards “Judicialization” and Uniformity?: Twelfth Sokol Colloquium (Irvington, NY: Transnational Publishers, 1994) 161.

[49] Pieter Sanders, “New York Convention on the Recognition and Enforcement of Foreign Arbitral Awards” (1994) 6:55 Nethl Intl L Rev 199.

[50] Supra note 38, at 20.

[51] Ibid.

[52] Supra note 22, at 313.

[53] Pierre Lastenouse, “Why Setting Aside an Arbitral Award is not Enough to Remove it from the International Scene” (1999) 16:2 J Intl Arb 25 at 43.

[54] Supra note 14.

[55] Article 1502 of the NCCP did not contain grounds to the effect of Article V(1)(e).

[56] Steven L. Smith, “Enforcement of International Arbitral Awards Under the New York Convention” in Rufus von Thulen Rhoades, Daniel M. Kolkey & Richard Chernick, eds, Practitioner’s Handbook on International Arbitration and Mediation (Huntington, NY: JurisNet, 2007) I.8.

[57] Supra note 38, at 212.

[58] Supra note 22, at 332.

[59] Hilton v Guyot, 159 U.S. 113, AT 227 (1895) (Supreme Court, US).

[60] Gary H. Sampliner, “Enforcement of Nullified Foreign Arbitral Awards – Chromalloy Revisited” (1997) 14:3 J Intl Arb 141 at 160.

[61] Supra note 43, at 197.

[62] Andrew Waters & Kimberley Smith, “Enforcement of Foreign Arbitration Awards”, Arbitration: What In-House Lawyers Need to Know, (King & Wood Mallesons, 2016).

[63] [2014] IEHC 115.

[64] [2013] DIFC 2.

[65] Cass civ 2°, 29 June 2007, (2007) Rev. de l’Arb 515 at [2].

[66] Supra note 22, at 333.

The PDF version of this article is available for download here.

Leaving an Empty Seat: Wong Souk Yee’s answer to By-Elections in a Group Representation Constituency

By Benjamin Low Junjie

I. Introduction

In Vellama d/o Marie Muthu v Attorney-General1 [Vellama], the Court of Appeal held that Article 49 of the Constitution of Singapore,2 which deals with the filling of vacant Parliamentary seats, imposed a duty on the Prime Minister to call a by-election “to fill casual vacancies of elected MPs which may arise from time to time”.3 However, despite the supposed universality of its conclusion, the Court of Appeal was quick to confine its judgment to vacancies arising in the context of Single-Member Constituencies (“SMCs”), having observed that separate legislative provisions and rules applied to instances where vacancies arose in a Group Representation Constituency (“GRC”).4 It remained an open question then as to what the proper scope and effect of Article 49 was on GRCs.

The recent High Court decision of Wong Souk Yee v Attorney-General5 [Wong Souk Yee] has since sought to address that question by holding that by-elections for a GRC would only be held in the event that all Members of Parliament (“MPs”) of that GRC have vacated their seats. This article will analyse the reasoning and the methodology of the High Court in reaching its conclusion and will discuss the implications of the judgment in the realm of constitutional interpretation.

II. Background

In Wong Souk Yee, the Applicant sought mandatory orders requiring that the MPs for the Marsiling-Yew Tee GRC vacate their seats in Parliament and that the Prime Minister advise the President to issue a writ of election for the Marsiling-Yew Tee GRC.6 Her action was founded on the basis that a vacancy had arisen in the Marsiling-Yew Tee GRC when one of its MPs, Madam Halimah Yacob, vacated her seat in Parliament on 7 August 2017 to contest in the 2017 Presidential Election.7 Since Madam Halimah’s departure, the Marsiling-Yew Tee GRC continued to be represented by its three remaining MPs from the People’s Action Party.8

The Applicant further sought declaratory orders in the alternative that s 24(2A) of the Parliamentary Elections Act9 [PEA] be interpreted as requiring all the MPs of a GRC to vacate their seats (i) when one or more MPs of the GRC vacate their seats; or (ii) in the alternative, when the only MP in the GRC who belongs to a minority community vacates his or her seat, so as to ensure consistency between s 24(2A) of the PEA and Article 49 of the Constitution.10 On the flip side, if such an interpretation was not possible, s 24(2A) was to be declared void for inconsistency with the Constitution as per Article 4 read with Article 49.11

III. The High Court’s Decision

A. The Mandatory Orders

The Applicant justified her claim for the mandatory orders on three grounds. Firstly, she contended that Article 49(1) of the Singapore Constitution 12 requires the calling of a by- election in a GRC when the seat of an Elected MP has become vacant for any reason other than the dissolution of Parliament. Secondly, the Applicant argued that Article 39A(2) of the Constitution13 requires that a GRC be represented by an MP belonging to a minority community, until the dissolution of Parliament. Last but not least, it was argued that a by-election should be called so as to give effect to the voters’ right to be represented by an elected MP of their choice.14

Much of the judicial debate in Wong Souk Yee centred on the Applicant’s first ground, which in effect concerned the proper interpretation of Article 49. For completeness, the full text of Article 49 reads as such:

Filling of vacancies

49.—(1) Whenever the seat of a Member, not being a non-constituency Member, has become vacant for any reason other than a dissolution of Parliament, the vacancy shall be filled by election in the manner provided by or under any law relating to Parliamentary elections for the time being in force.

(2) The Legislature may by law provide for —

(a) the vacating of a seat of a non-constituency Member in circumstances other than those specified in Article 46;

(b) the filling of vacancies of the seats of non-constituency Members where such vacancies are caused otherwise than by a dissolution of Parliament.

Article 49 in its present form was left untouched by the 1988 Amendments15 to the Constitution implementing the GRC system. Prior to that, the last substantive change to Article 49 was in 1984 concerning the introduction of Non-Constituency Members of Parliament.16 No other amendments were made to Article 49 since then. Indeed, when Article 49 was first enacted in the 1980 Reprint of the Constitution, 17 the GRC system was not even in existence. As Chua Lee Ming J succinctly identified, the true question that the High Court had to address, was the correct interpretation of Article 49(1) was in relation to GRCs.18 The earlier decision of Vellama had only addressed the scope of Article 49 in the context of an SMC.

The Applicant argued, on a plain reading of Article 49(1), that a by-election had to be called whenever a Parliamentary seat was rendered vacant, regardless of whether the seat belonged to an SMC or a GRC.19 Thus, whenever a vacancy arose in a GRC, the remaining MPs would have to resign their seats in order for a writ of election to be issued for the entire GRC. This would have to occur even if these MPs had no intention to vacate their respective seats, and where the conditions for disqualification or termination under Articles 45 and 46 were not made out.20 This interpretation of Article 49 was necessary in order to give substantive effect to the requirement in s 27A of the PEA 21 that elections in GRCs had to be held between teams of candidates.

The Attorney-General submitted in response that Article 49(1) only required that a by-election in a GRC be held if all the MPs of that GRC vacated their seats mid-term or, in the alternative, that Article 49(1) did not apply to GRCs but only to SMCs.22

In accepting the Attorney-General’s submissions, Chua J noted that the Applicant’s purported interpretation of Article 49(1) was a legal improbability given that there was no provision in Articles 45 and 46 of the Constitution requiring that the remaining MPs of a GRC had to vacate their seats or were disqualified from holding their seats in the event one of their fellow MPs vacated his or her seat.23 This is correct insofar as the Constitution is silent on such a contingency occurring whereas s 24(2A) of the PEA24 specifically mentions that the writ of election would only be issued in the event all MPs in a GRC have vacated their seats. To imply another constitutional ground on which an MP’s seat (or a group of MPs in this case) would either be vacated or the MP be disqualified from becoming a Member of Parliament would be tantamount to reading-in additional provisions in the Constitution by the Judiciary in the absence of clear or express words to that effect.25

Chua J further held in favour of the Attorney-General’s interpretation of Article 49(1) on the basis that either an updating or rectifying construction could be applied to Article 49(1) to take into account the subsequent legislative amendments to the Constitution and the Parliamentary Elections Act, thereby allowing for an interpretation of Article 49(1) to mean that by-elections in a GRC would only be possible if all MPs had vacated their seats. 26

An updating construction is first and foremost a rule of statutory interpretation that is meant to provide guidance “as to the legislator’s prima facie intention regarding the legal meaning of the enactment.”27 For an updating construction to take effect, the Act in question must first have been an Act “that is intended to develop in meaning with developing circumstances”.28 Once this has been established, the updating construction operates by presuming that “Parliament intends the court to apply to an ongoing Act a construction that continuously updates its wording to allow for changes since the Act was initially framed”.29 In Singapore, the rule of updating construction was first expressly adopted in Comptroller of Income Tax v MT 30 [CIT]. This was in turn cited in Wong Souk Yee with approval,31 as part of a three-stage framework (the “CIT Framework”).32

In applying the CIT Framework to the present case, Chua J determined that since Parliament expressly intended for by-elections in a GRC to be held only in the event that all seats in that GRC had been vacated33 and had made its intentions manifest by amending the Constitution to include Article 39A, an updating construction could be applied to Article 49(1) to reflect the changes effected by Article 39A.34

B. Difficulties with Applying an “Updating Construction”

With respect, the learned Judge did not appear to have adequately considered the nature of the legislation he was required to interpret in accepting the applicability of the doctrine of updating construction. As previously stated, the presumption of updating construction is a rule of statutory interpretation. It owes its development to the peculiar characteristics of the English legal system, where there exists no written Constitution and the doctrine of Parliamentary supremacy prevails.35 Given such differences between our legal systems, the Singapore courts ought to guard against resorting to ordinary canons of statutory interpretation too quickly in order to interpret a written constitution that proclaims itself to be the supreme law of the land. To infelicitously import rules of interpretation and construction for ordinary statutes wholesale into the realm of constitutional interpretation runs the risk of treating the Constitution as being no different from an ordinary piece of statute, even if such an interpretation is justified as giving effect to the purpose of the affected legislation.36

Even if one accepts that the rule of updating construction can be applied in the context of constitutional interpretation, it is arguable that at the third stage of the CIT Framework, the application of an updating construction to Article 49(1) would itself have entailed a substantive change to the provision. With respect, a constitutional amendment may have been the more desirable solution. As the Court of Appeal in Vellama so held, Article 49(1) in its present form mandates that the Prime Minister has to call by-elections “to fill casual vacancies of elected MPs that may arise from time to time”.37 This mandatory requirement is not subject to qualifications whatsoever on a plain reading of the text. Neither is it reflected in the 1988 Constitutional Amendment that Parliament intended for by-elections in GRCs to be subject to the requirement that all MPs vacate their seats first. Not even Article 39A itself makes any reference to the requirements for by-elections to GRCs, save that they be held only on a “team” basis.38 Furthermore, the PEA’s unique requirements that must be met for a by-election in a GRC are themselves a significant amendment to the existing constitutional provision on by-elections since the departure of one or more MPs from a GRC would not trigger the by-elections mechanism unless every MP has vacated his or her seat. The result is that in practice, Parliamentary vacancies in a GRC would be left unfilled until the remaining MPs’ terms of office ends.

That being the case, the proper remedy would lie with Parliament to amend Article 49 itself to accommodate s 24(2A) of the PEA. As Menon JC observed in CIT, where a “significant legislative change has been effected in the first statute such that its importation into the second statute may have quite substantial effects, the court should ordinarily be slow to apply an updating construction to the second statute”.39

C. Rectifying Construction: A Trojan Horse of Judicial Legislation?

A similar argument may be made about the Court’s willingness to apply a rectifying construction towards Article 49(1). Rectifying construction operates on the basis that “the legislator intends the court to apply a construction which rectifies any error in the drafting of the enactment”.40 Its effect is similar to that of an updating construction, and essentially involves the addition or substitution of words into the statute to rectify the error or omission in the legislative drafting. In accepting the Attorney-General’s argument, Chua J reasoned that Parliament’s intention regarding by-elections in GRCs was manifested by the adoption of s 24(2A) of the PEA.41 That Article 49(1) was not also amended to reflect this change in the law was merely an omission on the Legislature’s part which in no way detracted from the express intention of Parliament.42

With respect, it is still difficult to reconcile Chua J’s holding with the fact that the PEA is itself a piece of ordinary legislation that must inevitably rank below the Constitution in terms of legal hierarchy.43 Constitutional supremacy as reflected in Article 4 mandates that any ordinary legislation that would at first blush be inconsistent with the Constitution ought to rendered void to the extent of the inconsistency unless the Constitution itself has been expressly amended.44 S 24(2A) arguably constitutes a marked deviation from the constitutional duty in Article 49 to fill Parliamentary vacancies through the use of elections. It follows that an amendment to Article 49(1) is necessary to reflect this change, and in the absence of such, s 24(2A) is void for inconsistency qua Article 49(1).45

Having accepted the Attorney-General’s interpretation of Article 49(1), the High Court proceeded to consider the Attorney-General’s alternative interpretation that Article 49(1) only applied to SMCs and not to GRCs. In agreeing with the Attorney-General’s alternative interpretation, Chua J reiterated his earlier holding that the legislative purpose was clear in that:

“[A]ll elections (including by-elections) in any GRC shall be held on a basis of such number of candidates as designated for that constituency by the President, and no by- election needs to be held to fill any vacancy in a GRC unless all the Members in that GRC have vacated their seats.”46

The Attorney-General’s alternative interpretation of Article 49(1) was thus adopted as the correct one that “furthers the legislative purpose”.47 Yet it seems unclear whether the legislative purpose that the High Court was referring to was the specific purpose of Article 49(1) or the general purpose of Part VI of the Constitution or even the Constitution as a whole. If Chua J was referring to the specific purpose of Article 49(1), then insufficient consideration was devoted to ascertaining the legislative purpose of Part VI of the Constitution, as per the test in Tan Cheng Bock v Attorney-General48 [Tan Cheng Bock]. If, however, the learned Judge was referring to the legislative purpose of Part VI of the Constitution, his definition of the legislative purpose would appear to be an unduly narrow one. One must consider that Part VI deals wholly with the functions, organisation as well as composition of the Legislature of Singapore in its entirety, rather than simply on by-elections alone.49

Since the Applicant’s claim could not succeed on her first ground, the High Court proceeded to consider the remaining two grounds underpinning her claim for mandatory orders. However, as the Applicant clarified in her oral submissions, her case was only that a by-election in a GRC had to be called if any seat in the GRC was vacated regardless of whether it was held by an MP belonging to a minority community or otherwise. She thus conceded her second ground that Article 39A(2) of the Constitution requires that a GRC be represented by an MP belonging to a minority community until the dissolution of Parliament.50 The High Court further rejected her third ground of argument, that there existed an implied right by voters to be represented by an elected MP of their choice until the dissolution of Parliament.51

D. The Declaratory Orders

Given that the Applicant’s alternative claim for declaratory orders was founded on similar grounds as her claim for mandatory orders, the High Court rejected the Applicant’s alternative claim for the same reasons that it gave in rejecting the Applicant’s primary claim – namely that the Applicant’s proposed interpretation of Article 49(1) was patently incorrect. 52 Nevertheless, this author contends that, given the legal impossibility of the Applicant’s interpretation of Article 49(1) and the inherent risks of applying either an updating or rectifying construction to the Constitution, the irresistible conclusion appears to be that s 24(2A) of the PEA is simply inconsistent with the Constitution and thus void. It is unfortunate that the High Court did not canvass this particular point in greater detail, being content to simply reject it as stemming from the Applicant’s incorrect interpretation of Article 49(1).

IV. Implications for Constitutional Interpretation

Wong Souk Yee appears to reinforce the proposition that in future cases, the purposive approach53 will increasingly become the dominant method of constitutional interpretation.54 Coming on the back of the earlier decision of Tan Cheng Bock, which was itself heavily cited in Wong Souk Yee, it would seem that future cases dealing with questions of constitutional interpretation will inevitably concern themselves with attempting to ascertain the purpose of a particular Constitutional provision or a part of the Constitution. Yet, two potential issues have arguably arisen from the present case that necessitate further clarification from the Singapore courts.

Firstly, the apparent willingness of the High Court in Wong Souk Yee to resort to interpretive doctrines for ordinary legislation in interpreting the Constitution intimates their general applicability across different types of legislative instruments. It is true that Tan Cheng Bock itself appeared to accept the applicability of such rules in the context of constitutional interpretation.55 However, care must be taken against bluntly importing these rules into the separate realm of constitutional interpretation, especially since they were developed by the Courts in common law, and intended for ordinary statutes. A liberal use of such rules and canons runs the risk of the courts straying into the field of constitutional amendment under the guise of interpretation, especially if such rules have the practical effect of injecting additional words and expressions into the Constitution when none previously existed. Further clarification from the Court of Appeal as to the ambit and applicability of these rules would be greatly welcomed.

Secondly, this increasing acceptance of the purposive approach as the primary method of constitutional interpretation poses a corollary question: whether the alternative rights-based approach, as laid out in the long line of cases stretching from Minister of Home Affairs v Fisher 56 [Fisher] to Ong Ah Chuan v Public Prosecutor 57 [Ong Ah Chuan] and then to Taw Cheng Kong (which cited both Fisher and Ong Ah Chuan with approval), will continue to be of any particular relevance in field of constitutional law. It will be recalled that in Ong Ah Chuan, which cited Fisher and referred to the Singapore Constitution, the Privy Council held that “the way to interpret a Constitution on the Westminster model is to treat it not as if it were an Act of Parliament but “as sui generis, calling for principles of interpretation of its own, suitable to its character...without necessary acceptance of all the presumptions that are relevant to legislation of private law””.58

It is arguable that if the so-called sui generis approach in Ong Ah Chuan were applied to the present case, a different result might have been reached by the High Court with regards to the interpretation of Article 49(1). Presumably, the Applicant’s argument that there existed an implied right to representation within the Constitution would’ve been accorded more weight in determining the proper construction of Article 49(1).

While Ong Ah Chuan was decided before the purposive approach was expressly adopted as part of Singapore law, it remains good law nevertheless and presents a possible alternative approach towards constitutional interpretation. However, it will be necessary to re-examine whether the so-called sui generis approach in Ong Ah Chuan ought to remain part of Singapore case law in light of the increasing resort to the purposive approach in constitutional interpretation.

V. Conclusion

Wong Souk Yee sought to address the latent ambiguity surrounding the scope of the by-elections clause in Article 49 of the Constitution vis-à-vis the GRC scheme by holding that the provision could be updated or rectified to incorporate changes made elsewhere to existing electoral law. While the High Court’s conclusion appears to be technically correct based on the express wording of existing legislation regarding Parliamentary elections as well as an unduly broad application of the purposive approach in interpreting Article 49(1), the reasoning of the High Court has arguably opened a can of worms regarding the proper interpretation of constitutional provisions. The decision also runs a noticeable risk of undermining the very notion of constitutional supremacy. While it remains to be seen whether an Appeal will be brought, further clarification from the Court of Appeal as to the abovementioned points of contention would be greatly welcome.

[1] [2013] 4 SLR 1; [2013] SGCA 39.

[2] Constitution of the Republic of Singapore (1999 Rev Ed), art 49.

[3] Supra note 1 at [82].

[4] Ibid at [80] and [82].

[5] [2018] SGHC 80.

[6] Ibid at [3].

[7] Royston Sim, “Halimah Yacob steps down as Speaker and MP, resigns from PAP to run for President”, The Straits Times (7 August 2017), online: <https://www.straitstimes.com/politics/halimah-yacob-steps-down-as-speaker-and-mp-resigns-from-pap-to-run-for-president>.

[8] Valerie Koh & Kenneth Cheng, “3 remaining Marsiling-Yew Tee GRC MPs to build on Halimah’s good work”, TODAYonline (7 August 2017), online: <https://www.todayonline.com/singapore/3-remaining-grc-mps-build-halimahs-good-work>.

[9] Cap 218, 2011 Rev Ed Sing.

[10] Supra note 5 at [4].

[11] Ibid.

[12] Supra note 2 at art 49(1).

[13] Ibid at art 39A(2).

[14] Supra note 5 at [12].

[15] Constitution of the Republic of Singapore (Amendment) Act, No 3 of 1988, amending Constitution of the Republic of Singapore (1985 Rev Ed).

[16] Constitution of the Republic of Singapore (Amendment) Act, No 16 of 1984, s 5, amending Constitution of the Republic of Singapore (1980 Reprint), Art 49.

[17] Constitution of the Republic of Singapore (1980 Reprint).

[18] Supra note 5 at [9].

[19] Ibid at [14].

[20] Supra note 2 at arts 45 and 46.

[21] Supra note 9 at s 27A.

[22] Supra note 5 at [16]–[18].

[23] Ibid at [23]–[26].

[24] Supra note 21 at s 24(2A).

[25] On the dangers of unduly expanding the scope of or implying into the Constitution additional interpretations, see Rajeevan Edakalavan v Public Prosecutor [1998] 1 SLR(R) 10; [1998] SGHC 2 at [19].

[26] Supra note 5 at [27].

[27] FAR Bennion & Oliver Jones, ed, Bennion on Statutory Interpretation: A Code, 6th ed (London, UK: LexisNexis, 2013) at 797.

[28] Ibid.

[29] Ibid.

[30] [2006] 3 SLR(R) 688; [2006] SGHC 120.

[31] Supra note 5 at [28]–[29].

[32] Supra note 30 at [46]–[47].

[33] Supra note 5 at [36]; see also Parliamentary Debates Singapore: Official Report, vol 50 at cols 334–335 (12 January 1988) (Mr Goh Chok Tong).

[34] Ibid at [33]–[38].

[35] As expressly acknowledged in Bennion, supra note 27 at 167.

[36] See VK Rajah, SC, “Interpreting the Singapore Constitution” in Jacyln L Neo, ed, Constitutional Interpretation in Singapore: Theory and Practice (New York: Routledge, 2017) 23 at 24.

[37] Supra note 1 at [82].

[38] Supra note 2 at art 39A(2)(c).

[39] Supra note 30 at [48].

[40] Supra note 27 at 788.

[41] Supra note 5 at [41].

[42] Ibid.

[43] Halsbury’s Laws of Singapore, vol 1 (Singapore: LexisNexis, 2017 Reissue) at 100, para 10.506.

[44] Supra note 2 at Art 4. See also Chan Hiang Leng Colin v Public Prosecutor [1994] 3 SLR(R) 209; [1994] SGHC 207 at [50] and Taw Cheng Kong v Public Prosecutor [1998] 1 SLR(R) 78; [1998] SGHC 10 [Taw Cheng Kong] at [14] for judicial recognition of the doctrine of constitutional supremacy.

[45] While no statute passed by Parliament has ultimately been held unconstitutional by the courts, there nevertheless exists judicial precedent for such an approach: see Taw Cheng Kong, ibid, at [84].

[46] Supra note 5 at [47].

[47] Ibid at [49].

[48] Tan Cheng Bock v Attorney-General [2017] 2 SLR 850; [2017] SGCA 50 at [41].

[49] This can be contrasted with Vellama, supra note 1 at [79], where the Court of Appeal interpreted Article 49(1) with reference to the broad purpose of the Constitution in establishing a Westminster model of government with a right to representation.

[50] Supra note 5 at [52]–[53].

[51] Ibid at [57].

[52] Ibid at [59]–[60].

[53] Interpretation Act (Cap 1, 2002 Rev Ed), s 9A; see also Constitution of the Republic of Singapore (1999 Rev Ed) at art 2.

[54] Some commentators have welcomed this unified approach for both constitutional and statutory interpretation: see Goh Yihan, “The Interpretation of the Singapore Constitution: Towards a unified approach to interpreting legal documents” in Jaclyn L Neo, supra note 36, 257 at 276.

[55] Supra note 48 at [38].

[56] [1980] AC 319.

[57] [1979-1980] SLR(R) 710; [1980] SGPC 6.

[58] Ibid at [23].

The PDF version of this article is available for download here.

Considerations in an Autonomous Era

by Keith Wong

Enter the autonomous age. Once the far-flung conception of a childhood storybook, the entry of autonomous technology into our daily lives is now becoming a common reality. Amid the dazzle of such innovative wonder, prudence warrants that we consider how to manage liability and ethics when things turn awry. With increasing adoption of autonomous technology, is the law capable of contending with a shift to this age? In this article, an assessment of potential manufacturer’s liability will be examined.

The accepted taxonomy for the classification of autonomous vehicles has been categorized into six different stages.1 This is outlined in the diagram below:2

 As reproduced from Mr. Brian Buntz's article on the Internet of Things Institute. All rights in this photograph belong to him.

As reproduced from Mr. Brian Buntz's article on the Internet of Things Institute. All rights in this photograph belong to him.

At the highest level, vehicles operate at full automation, with limited to no driver intervention required in moving from point A to point B. Most vehicles presently on the road function at the lower end of this operating spectrum instead, requiring significant driver intervention.

However, an increasing range of vehicles available for sale have crept up the spectrum, offering assistance to drivers at Levels 1 and 2. For example, some vehicles are able to rely on radar sensors to navigate into parallel parking lots, while others offer adaptive cruise control and automatic emergency braking.

In the past few months, videos of vehicles equipped with highly autonomous functions have surfaced. They depict consumers, who purchased these vehicles from a particular American manufacturer, taking leisurely naps at the wheel, or in some cases leaving the driver’s seat vacant while the car is in operation. It invariably begs the question – what happens when an accident occurs?

In 2016, a nuTonomy driverless vehicle collided into a lorry at one of the firm’s testing grounds. This was headlined as the first accident in Singapore involving an autonomous vehicle. In a report, the Straits Times pointed to software glitches which affected how the vehicle detected and responded to other vehicles in the vicinity.3 As this issue was subsequently resolved, there was no need for a case to be raised.

Given the complex nature of such technology and its current, limited commercial availability, such a concern has yet to be highlighted by our local courts. In due time however, a number of considerations are likely to arise in regard to this field.


From a consumer’s standpoint, a selling point for autonomous technology lies in the data collected by the manufacture and the software developed for each product. What distinguishes one manufacturer’s product from the next is the sophistication and unique qualities offered. In other words, an autonomous vehicle is only as good as the manufacturer has programmed it to be.

The concept of a duty of care can be traced back to Donoghue v Stevenson4 [Donoghue], as well as the earlier case of Heaven v Pender5 [Heaven]. The cases themselves are black letter, but what is significant in Heaven lies in Brett MR’s assessment6 that established an obligation between persons in the absence of a contract between them. Lord Atkin built upon this proposition in Donoghue with the notion of the neighbour principle. This cemented the existence of a duty of care between a manufacturer and a consumer, premised on the concept of persons who are so closely and directly affected by the act that the manufacturer ought reasonably to have them in contemplation. Clearly, the same principles are likely to apply to the relationship between manufacturers and purchasers of autonomous vehicles.

In Singapore, the landmark decision of Spandeck Engineering v Defence Science & Technology Agency7 establishes the local approach in determining whether a duty of care exists. At the threshold stage, factual foreseeability of harm arising from the defendant’s negligence is first evaluated. This is followed by an assessment of proximity and nexus between parties. In the subsequent case of Anwar Patrick Adrian v Ng Chong & Hue LLC8, the Court expanded the range of proximity factors to recognize knowledge, control, vulnerability, assumption of responsibility and reliance, as suggested by Professors David Tan and Goh Yihan.9 Finally, policy concerns have a determinative role in whether the imposition of a duty of care is negated.

In assessing the local position on a duty of care in relation to a consumer’s claim against a manufacturer, the presence of a proximate relationship is likely uncontentious. However, the position on policy is likely to be probative with reasons both in favour of and against imposing a duty of care on manufacturers of autonomous vehicles.

At present, Singapore takes a favourable position toward research and innovation. This is reflected by strong investment in innovation from both the public and private sector, such that “Singapore’s [Research and Development] efforts have led it to be consistently ranked in the top 10 in the Global Innovation Index.”10 Given this stance, policy may negate the imposition of a duty of care on manufacturers, in order to encourage innovation in a sandbox environment.

However, a comparative examination illustrates a variance in legislative attitudes toward liabilities arising from autonomous vehicles. Europe has launched several initiatives relating to Connected and Automated Driving.11 In Belgium, liability on the part of the car manufacturer may be civil or even criminal, such that both the company and its directors may be prosecuted.12 A contrasting approach to liability has been taken on the other side of the globe. Recent amendments to the California Department of Motor Vehicles removed Section 227.38(b), which originally read:

“The manufacturer certifies that, to the extent the manufacturer’s autonomous technology causes the autonomous vehicle to be at fault in a collision, the manufacturer shall assume liability for damages caused by the autonomous vehicle in such collision, but subject to applicable law.”13

The subsequent amendment to excise part of this clause suggests that existing liability laws in California will apply instead, as no indication is given as to whether manufacturers assume liability for damages caused by autonomous vehicles. However, it is key to note that this regulation applies to driverless testing and public use rules for developing autonomous vehicles. As the focus is on the testing and development stage, it is thus appropriate to consider what may apply to products already available for purchase. Consequently, the lack of a global standard and its various ambiguities, raises the need to consider what may develop in our local environment.

Even with a duty present, the standard of care expected of makers of autonomous vehicles is also open at law. Typically, the standard of care is an objective assessment of what a reasonable person would do in the circumstance. This applies well in cases of negligence arising from human error or oversight. However, the multiplicity of benchmarks for the performance of autonomous vehicles suggests that unanimity on the applicable standard is yet settled.

Across the board, imposing a high standard of care for operators of vehicles has been a relevant policy consideration. From the traditional position reflected in Nettleship v Weston14 to the more recent case of Ng Keng Yong v PP,15 factors such as inexperience have failed to reduce the requisite standard of care. In Singapore, this posture has been further emphasized in Public Prosecutor v Hue An Li16. In his written judgment, Sundaresh Menon CJ reiterated the high standard of care expected,17 stressing determinedly that “[t]he starting point for sentencing in a s 304A(b) traffic death case is a brief period of incarceration for up to four weeks.”

Admittedly, the lack of global unanimity begs the question as to what standard manufacturers of autonomous vehicles should be subject to. Manufacturers might prefer a subjective standard as it allows for a more flexible approach to product development, but this tugs against the generally-accepted benchmark of an objective, reasonable manufacturer. Even this objective benchmark is left open for determination: while features such as reliability and functionality would generally be expected of traditional automobiles in the eyes of consumers, there is no global industry standard for autonomous vehicles on this front. This is likely to span a host of future dispute between policy makers and manufacturers. It probes at an incipient tension between protecting consumers by way of a higher objective standard, against a more relaxed approach which encourages innovation. As suggested by this article, one solution to address such ambiguity in the law, may be to draft legislation tailored to address such concerns.


A duty to take care in programming is likely to arise alongside other implied contractual warranties provided for by legislation in the Sale of Goods Act.18 Attempts to hold manufacturers liable will also likely be met with attempts to reduce or exclude liabilities, which also raise concerns as to whether statutes such as the Unfair Contract Terms Act19 [UCTA] apply. For example, dealers with Level 1 and Level 2-type vehicles for sale often require buyers to sign an exclusion of liability form, absolving them from liability in the event that the technology fails to act as it should. In determining whether a term of the contract has been breached, it is also appropriate to look for any warranties given by the manufacturer on how their product is meant to operate. Should negligent programming on the part of the manufacturer result in the causing of death or personal injury however, UCTA, as well as the contra proferentem rules suggest that such liability cannot be indemnified against.


What then happens, when a driverless vehicle operating at the higher end of the autonomous spectrum meets with an accident due to its own erroneous calculations?

From a manufacturer’s point of view, this raises a possible defence of novus actus interveniens on the part of the driver in the law of tort. It is, for instance, at least arguable that a consumer’s recklessness by sleeping in the driver’s seat while the vehicle functions in autonomous mode on a congested expressway is so wholly unreasonable that it amounts to a novus actus interveniens. Admittedly, this is a slippery slope, given that the trajectory of autonomous vehicles aims to be less dependent on human intervention. Hence, this area of the law is certainly one to watch, given the impending arrival of this technology and the inevitability of such a claim.


Shifting back to the scene of an accident, matters become more complicated when the circumstance develops to value human life. How would an autonomous vehicle contend with ethical choices, such as the dilemma of saving many at the expense of the few? In response to this trolley problem,20 Professor Reich has questioned whether autonomous vehicles will “[o]ptimize for overall human welfare” when faced with the need to prioritize one human life over the other.21

Separately, this raises an interesting question in relation to the provisions of the Penal Code.22 As indicated by the preliminary section of the wording and scope of the Penal Code, the provisions criminalize the acts of natural and legal persons. In an extreme scenario that an autonomous vehicle runs over a pedestrian, applying many provisions of the Penal Code would stretch the realm of reality. This prompts for future development of the Penal Code, and encourages the drafting of more statutes pertaining to advanced technology.

Outside of the Penal Code, one such example has been the development of the Cybersecurity Bill,23 in response to concerns brought about by the Fourth Industrial Revolution. As technology continues to disrupt the traditional world we know, the spill-over effects into the legal system demands changes that keep pace with the steep learning curve. This is especially in light of the progressively symbiotic relationship that autonomous technology and artificial intelligence will share with human beings.


One response to increasing developments in this field has been in regulation via the Road Traffic Amendment in 2017.24 This amendment addressed the testing of autonomous vehicles in Singapore, and paves the path toward future regulation and development governing the operation of autonomous technology. With Singapore poised as a technologically adept society, continued emphasis in this area of the law can encourage local growth in this field, while placing Singapore as an attractive hub for future innovation.


While this article merely scratches the tip of the iceberg, such concerns are likely to become increasingly prevalent. With the rise of autonomous technology and a growing global investment in artificial intelligence, the time is ripe to re-assess traditional areas of the law. On the cusp of the Fourth Industrial Revolution, the opportunities and need to be familiar with changing technology are certainly pressing. Appropriately, this encourages lawyers (and lawyers-to-be) to address and to understand the growing demands in this area.

[1] On-Road Automated Driving Committee, “Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles”, SAE International (30 Sep 2016), online: < http://standards.sae.org/j3016_201609/> (accessed 2 Jan 2018).

[2] Brian Buntz, “What are the 5 Levels of Autonomous Driving?”, Internet of Things Institute (22 Aug 2016), online: <http://www.ioti.com/transportation/what-are-5-levels-autonomous-driving> (accessed 2 Jan 2018).

[3] Zhaki Abdullah, “‘Software Anomalies’ to Blame for Driverless Car Accident”, The Straits Times (25 Nov 2016), online: <http://www.straitstimes.com/singapore/software-anomalies-to-blame-for-driverless-car-accident> (accessed 2 January 2018).

[4] M’Alister (or Donoghue) (Pauper) v Stevenson [1932] AC 562.

[5] (1883) 11 QBD 503.

[6] Ibid, at p 507.

[7] [2007] 4 SLR(R) 100.

[8] [2014] 3 SLR 761.

[9] In “The Promise of Universality: The Spandeck Formulation Half A Decade On” [2013] 25 SAcLJ 510.

[10] Lim Chuan Poh, “From Research to Innovation to Enterprise: The Case of Singapore” in Soumitra Dutta, Bruno Lanvin, and Sacha Wunsch-Vincent, eds, The Global Innovation Index 2016 (World Intellectual Property Organization) 133 at 138.

[11] Matthieu Relange, “At a Glance: Autonomous Vehicles, EU”, Bird & Bird (25 July 2017) < https://www.twobirds.com/en/news/articles/2017/global/at-a-glance-autonomous-vehicles> (accessed 04 March 2018)

[12] Jaward Aadel, “At a Glance: Autonomous Vehicles, Belgium”, Bird & Bird (25 July 2017) < https://www.twobirds.com/en/news/articles/2017/global/at-a-glance-autonomous-vehicles> (accessed 04 March 2018)

[13] Cal. Code Regs. tit. 13, § 227.38(b), as indicated in the Second Modified Express Terms (California: Department of Motor Vehicles) at p 11, online: <https://www.dmv.ca.gov/portal/wcm/connect/aa08dc20-5980-4021-a2b2-c8dec326216b/AV_Second15Day_Notice_Express_Terms.pdf?MOD=AJPERES> (accessed 4 March 2018)

[14] [1971] 2 QB 691.

[15] Ng Keng Yong v Public Prosecutor and another appeal [2004] 4 SLR(R) 89.

[16] [2014] 4 SLR 661.

[17] Ibid, at [61].

[18] Cap 393, Rev Ed 1999.

[19] Cap 396, Rev Ed 1994.

[20] As raised by Professor Taylor Reich in Alex Shashkevich, “Stanford Scholars, Researchers Discuss Key Ethical Questions Self-Driving Cars Present”, Stanford University (22 May 2017), online: <https://news.stanford.edu/2017/05/22/stanford-scholars-researchers-discuss-key-ethical-questions-self-driving-cars-present/> (accessed 2 January 2018)

[21] Ibid.

[22] Cap 224, Rev Ed 2008.

[23] Bill No. 2 of 2018.

[24] Road Traffic Act (Cap 276, 2004 Rev Ed Sing), as amended by the Road Traffic (Amendment) Act No. 10 of 2017.

The PDF version of this article can be found here.

Being Seaworthy in Singapore: A brief critique of Article 14 of the Rotterdam Rules

By Tan Kah Wai


It was Lord Diplock who once described the seaworthiness obligation as one of the “most complex of contractual undertakings”.1 If the waters of seaworthiness were already not rocky enough, Art 14 of the United Nations Convention on Contracts for the International Carriage of Goods Wholly or Partly by Sea2 [Rotterdam Rules] just complicates matters further by extending this nebulous obligation throughout the voyage itself. This brief paper seeks to elucidate the legal implications that Art 14 of the Rotterdam Rules has on the seaworthiness obligation in shipping law. Particular regard is given to discussing the adverse effect on Singapore’s position as an international maritime hub should she choose to adopt the rules.3

For ease of reference, Article 14 of the Rules is set out as follows:

“The carrier is bound before, at the beginning of, and during the voyage by sea to exercise due diligence to:

(a) make and keep the ship seaworthy;

(b) Properly crew, equip and supply the ship and keep the ship so crewed, equipped and supplied throughout the voyage and

(c) Make and keep holds and all other parts of the ship in which the goods are carried, and any containers supplied by the carrier in or upon which the goods are carried, fit and safe for their reception, carriage and preservation.”


First, the carrier has an absolute obligation in common law to keep the ship seaworthy at the commencement of the voyage.4 However if the carriage contract is subject to Art 3.1 of the Hague-Visby Rules5 [HVR], the carrier is only bound to make the ship seaworthy by exercising due diligence. Art 4.1 of the HVR also makes it clear that carriers cannot contract out of this overriding and non-delegable obligation. These qualifications constitute the local position by virtue of the HVR’s inclusion under the Schedule of the Carriage of Goods by Sea Act.6

The duty of seaworthiness encompasses the following aspects:

  1. Ensuring that the vessel is structurally fit for the intended voyage;7
  2. Providing sufficient manpower and equipment, and ensuring that the vessel’s crew have an acceptable level of competence;8
  3. Providing relevant documentation and charts,9 especially those concerning health and safety;
  4. Ensuring that the ship is reasonably fit to receive and carry the cargo.10

In determining seaworthiness, the Singapore courts have consistently held that it is a relative standard that varies according to the ship and exigencies of the voyage.11 A seaworthy vessel is one that is as fit as ordinary owners would expect to have when the vessel commences its voyage. For the courts, this is an objective test that asks whether a prudent owner having known of the defect, would require it to be remedied before the ship is sent to sea.12 The burden of proof for seaworthiness first rests on the claimants.13 However where claimants can prove that there is damage or loss, a presumption of unseaworthiness may arise on the facts. The evidential burden thus shifts to the carrier to prove that he did his due diligence in keeping the ship seaworthy.14


The Rotterdam Rules were enacted to achieve a few aims:

  1. Updating of international carriage rules to align it with technological developments;
  2. Harmonization of international trade, in particular, achieving broad uniformity in the law governing international carriage of goods; and
  3. Balancing both commercial and national interests.15

It is in this spirit that the drafters hoped to adopt an appropriate legal framework for carrier’s duties that the shipping industry would find acceptable.16

At first glance, Art 14 of the Rotterdam Rules largely replicates HVR’s Art 3.1. However, by adding the words “and during the voyage”, it extends the duty of keeping the ship seaworthy to one that is throughout the voyage. The courts thus now have to ask an additional question: Would a prudent shipowner being aware of the defect have continued this voyage without undertaking any possible repairs?17

Under the HVR, the seaworthiness obligation was limited to before the voyage’s commencement. The assumption back then was that carriers had no control over the vessel once it has sailed. These assumptions are no longer relevant clearly since there is satellite technology that allows permanent communication between vessels on the sea and their offices at the shore.18 By extending the obligation of seaworthiness, Art 14 of the Rotterdam Rules reflects a change in modern circumstance where carriers now have a greater level of control over ships after commencement of voyage.19

Another justification for imposing this continuous obligation is that it would be consistent with the carrier’s public obligations under Articles 6 and 10 of the International Safety Management Code, which stipulate that the ship must be crewed, equipped and maintained properly throughout the voyage.

Yet one must ask whether there is a purpose undergirding the drive for consistency here. If achieving consistency is to ensure compliance with public obligations, then it is strictly unnecessary since this compliance can be very well achieved by regular inspections from the public authorities. Under Singapore’s Merchant Shipping Act,21 the Maritime Port Authority (“MPA”) of Singapore’s surveyors can undertake Flag State Control Inspections, which allows them to inspect and verify the ship’s safety management systems.22 The MPA is also entitled to arrest ships for any non-compliance with safety regulations.23

A third justification rests on the fact that the concept of continuous obligation is already part of commercial practice.24 For example, most time charterparties contain clauses like the NYPE 93 Form Clause 6, which stipulates a continuous obligation to provide a seaworthy vessel.25

However, this justification does not stand as it ignores the fact that this additional clause is often omitted in voyage charterparties for good commercial reasons. Shipowners in time charterparties enjoy security of income throughout the charter period since charterers are responsible for finding the ship’s commercial employment and payment of hire.26 However in voyage charterparties, the carrier pays virtually everything except for delays at loading and discharge ports. If Art 14 extends this continuous obligation to voyage charterparties, it imposes far more onerous obligations on carriers who must now undertake extra expenses to ensure the ship’s seaworthiness throughout the voyage.


A. It does not indicate when the Obligation of Seaworthiness ends

Firstly, Art 14 imposes a continuous obligation to keep the ship seaworthy without indicating to parties when this duty ends. The Rules only state that this duty persists “during the voyage” but remains silent on when does the voyage end.

Hence, Aladwani raises the following scenarios where the voyage may end:27

  1. When it enters the port’s geographical and legal area;
  2. When it anchors at a place where vessels lie while waiting for berth at a port; or
  3. Discharge of the cargo from the vessel itself. For a set of international rules that seek to introduce obligations unique to the common law, it is difficult to see why Art 14 fails to define a precise point at which the seaworthiness obligation would end. This is clearly unsatisfactory because it diminishes commercial certainty, which is acknowledged by the courts as an important consideration in international trade.28

B. It generates problems of evidentiary proof

Secondly, Art 14 presents practical problems in ensuring a continuous obligation for seaworthiness. This is especially when ships experience problems in the middle of the voyage, which is a frequent occurrence. While the drafters have acknowledged these concerns, they thought that such concerns would be resolved by stating that the obligation is fulfilled once a standard of due diligence is satisfied.29

However, just simply stating a due diligence standard does not remove the carriers’ evidential difficulties. Carriers still bear the burden of proving that they fulfilled their due diligence obligation of keeping the ship seaworthy throughout the voyage.

To illustrate these difficulties, consider the case where a ship sinks without collision or inclement weather. A res ipsa loquitur inference would be drawn that the ship is unseaworthy during the voyage.30 The carrier bears the burden of disproving this presumption. Yet, it cannot rely on the surveyors’ certification of the ship’s safety before the voyage. This is because claimants will frame the duty breached here as one where the ship is not seaworthy during the voyage, rather than it being unseaworthy before the voyage commences. Even if the carrier claims that the ship is unseaworthy due to a latent defect, it is still difficult to prove such defects where the ship has been wrecked.

C. Complying with Art 14 breaches other articles of the Rotterdam Rules

An example here would suffice to illustrate this point. A Carrier that seeks to repair the ship during the voyage may delay the cargo delivery in doing so. He has two options. If he chooses to delay the carriage, he may face liability under Art 21 of the Rules. On the other hand, he faces liability under Art 14 if he chooses not to remedy the defect. As Tsimplis has observed, the Carrier is ensnared in an absurd Catch-22 situation where it is liable either way.31

D. Art 14 has negative practical implications on all commercial parties in the shipping industry

Ironically, Art 14 can hardly be justified as one that enhances cargo interests in light of its aforementioned risks and commercial reality. the commercial realities would defeat any justification of Art 14 as one that enhances cargo interests. Carriers will take potentially cost- inefficient measures to ensure the ship’s seaworthiness, which translates to higher freight costs for shippers.32

While academics have assumed that the standards for due diligence under Art 14 is different at port than at sea,33 such assumptions provide scant comfort for risk-averse insurers.34 The travaux préparatoires as an interpretive tool for Art 14 is incapable of assisting us in determining these standards.35 In response to such heightened exposure in liability and risk of protracted litigation over what constitutes a breach of Art 14, insurers would likely raise insurance premiums for carriers. Carriers would, in turn, transfer these costs to shippers through higher freight rates.36 The result is far more commercially undesirable – a chilling of shippers’ demand for carriage services in a bearish market. Just last year, it has been observed that the fleet supply is already far excess of what has been sluggish demand, with liner companies being forced to actively restructure themselves to steady their costs.37

While this snowball effect has yet to be empirically proven, it is clear that insurers and carriers would prefer the pre-Rotterdam position. There were no obligations to keep the ship seaworthy throughout the voyage, having to resolve this legal question before the courts.

E. It impedes Singapore’s growth as a Maritime Legal Services Hub

Although shipping lines have expressed support for the Rotterdam Rules,38 it would be naive to assume that their support is unequivocal when it goes against their own commercial interests.

Shippers and carriers do engage in forum shopping to obtain a limitation of liability regime that is most advantageous to their case.39 Hence, it makes commercial sense for carriers to prefer forums where the HVR applies, rather than forums adopting Art 14 of the Rotterdam Rules that impose this continuing and uncertain obligation. This explains why none of the top 5 shipping centres in the 2017 Xinhua-Baltic International Shipping Centre Development Index (“Baltic Exchange Index”) including Singapore have taken the plunge and signed the Rotterdam Rules. Nations are refusing to ratify the rules until major trading nations have done so40 precisely because they fear a loss of business due to forum shopping.

Furthermore, despite Singapore being the top shipping centre in the Baltic Exchange Index, she only enjoys a minor share of the international maritime legal services market.41 There is little merit in adopting Art 14 now, where it may unnecessarily impede Singapore’s competitiveness and growth as a maritime legal services hub.


This paper has sought to demonstrate how Art 14 is functionally incoherent, lacks sensible commercial justification, and may even jeopardize Singapore’s maritime industry. The local maritime industry is already facing some challenging times ahead. Hence, it would be prudent for Singapore to consider carefully if she ultimately wishes to adopt the Rotterdam Rules.

[1] Hongkong Fir Shipping Co Ltd v Kawasaki Kisen Kaisha [1962] 2 QB 26, at 71

[2] 23 September 2009, A/RES/63/122

[3] The Rotterdam Rules have been controversial in many aspects, but this paper will focus solely on Art 14 for the sake of brevity.

[4] Sunlight Mercantile Pte Ltd and another v Ever Lucky Shipping Co Ltd [2003] SGCA 47, [12].

[5] International Convention for the Unification of Certain Rules of Law relating to Bills of Lading, 25 August 1924, 120 LNTS 187 (entered into force 2 June 1931)

[6] Cap 33, 1998 Rev Ed Sing.

[7] Zuellig (Gold Coin Mills) v Autoly (Owners) [1968-1970] SLR(R) 829; [1970] SGHC 17.

[8] Tan Lee Meng, The Law in Singapore on Carriage of Goods by Sea, 2nd ed (Butterworths, 1994) at 70-71.

[9] Alfred C Toepfer Schiffahrtsgesselschaft GmbH v Tossa Marine Co [1985] 2 Lloyd’s Rep 325, at 331.

[10] McFadden v Blue Star Line [1905] 1 KB 697, at 704.

[11] Malayan Motor & General Underwriter (Pte) Ltd v MH Almojil [1981-1982] SLR(R) 432; [1982] SGCA 3, at [26].

[12] Ever Lucky Shipping Co Ltd v Sunlight Mercantile Pte Ltd [2003] SGHC 80.

[13] The “Reunion” [1983-1984] SLR(R) 141; [1983] SGHC 12, at [30].

[14] The “Patraikos 2” [2002] SGHC 103, [133].

[15] See generally Sturley, “General Principles of Transport Law and the Rotterdam Rules” in Guner-Ozbek (ed), The United Nations Convention on Contracts for the International Carriage of Goods Wholly or Partly by Sea: An Appraisal of the “Rotterdam Rules (Springer, 2011) at 63-85.

[16] Nikaki, The Carrier’s Duties Under the Rotterdam Rules: Better the Devil You Know? (2010) 35:1, Tul Mar LJ 13.

[17] Nicholas, “The Duties of Carriers under the Conventions: Care and Seaworthiness” in D Rhidian Thomas (ed), The Carriage of Goods by Sea under the Rotterdam Rules (Lloyd’s List, 2010) at 6.14.

[18] Schoenbaum and Yiannapoulos, Admiralty and Maritime Law: Cases and Materials (The Michie Company, 1984) at 14-15.

[19] Nikaki and Soyer, “A New International Regime for Carriage of Goods by Sea: Contemporary, Certain, Inclusive and efficient, or just another one for the shelves?” (2012) 30 BJIL 303 at 329.

[20] United Nations Commission on International Trade Law [UNCITRAL], Report of Working Group (Transport Law) on the work of its twelfth session (Vienna, 6-17 October 2003), at [149].

[21] Merchant Shipping Act (Cap 179, 1996 Rev Ed Sing).

[22] Ibid, s 205.

[23] Ibid, s 207.

[24] Defossez, “Seaworthiness: The Adequacy of the Rotterdam Rules Approach”, (2015-2016) 28:2 USF Mar LJ, at 247.

[25] Girvin, Carriage of Goods by Sea, 2nd ed (Oxford University Press, 2011), at 24.25.

[26] Ibid, at 1.45.

[27] Aladwani, The Supply of Containers and “Seaworthiness – The Rotterdam Rules Perspective, 42:2 J Mar L & Com Journal of Maritime Law & Commerce 185 at 200.

[28] See for instance, Homburg Houtimport BV v Agrosin Private Ltd [2004] 1 AC 715, at [13].

[29] Supra note 18.

[30] A similar inference was raised in Ajum Goolam Hossen & Co v Union Marine Insurance [1901] AC 362.

[31] Tsimplis, “Obligations of the Carrier” in Baatz, Debattista et al (eds), The Rotterdam Rules: A Practical Annotation (Informa Law, 2009) at 14-06.

[32] Supra note 25 at 27.61.

[33] Supra note 31 at 14-03.

[34] Efthymiou, “Speech delivered by the President of the Union of Greek Shipowners” in CMI Yearbook 2009 Annuaire, Athens II, Documents of the Conference (CMI, 2009) at p 293.

[35] For it to be useful, it must clearly and indisputably point to a definite legal intention. See Fothergill v Monarch Airlines Ltd [1981] AC 251, at 278C.

[36] Supra note 34 at p 294.

[37] Baltic Exchange, “Xinhua-Baltic International Shipping Centre Development Index 2017” (Xinhua-Baltic Exchange, 2017) at 10.

[38] Supra note 15 at pp 76-77.

[39] Hare, “Shopping for the Best Admiralty Bargain” in Martin Davies (ed.), Jurisdiction and Forum Selection in International Maritime Law: Essays in Honour of Robert Force, (Kluwer, 2005) at p 163.

[40] Sturley, “The Rotterdam Rules in Beijing” in CMI Yearbook 2013 Annuaire, Beijing II, Documents of the Conference (CMI, 2013) at p 266.

[41] Supra note 37 at p 61.

The PDF version of this article is available here.

From Tragedy to Statistic: How Big Data has Changed the Practice of Law

by Bryont Chin

Data is information, and information never stops growing. In 2012, 2.5 exabytes1 of data were generated every single day.2 90% of the data in the world today has been created in the past two years.3 In Singapore, there is currently scant regulation pertaining to cybersecurity and data use. The Personal Data Protection Act4 and the Computer Misuse and Cybersecurity Act5 are the Legislature’s responses to these issues. These are mainly focused on preventing and penalizing cybercrime, especially intrusions into important government and private networks. However, while hacking is undoubtedly a perennial concern, in recent years many have asked another important question: how should the law respond to the increasing importance of big data and data analytics?

The term “big data” has been variously defined, but it centrally refers to “extremely large data sets that may be analysed computationally to reveal patterns, trends, and associations, especially relating to human behaviour and interactions”.6 Multinational companies and governments can collect these data sets, analysing them to generate insights which cannot typically be gleaned through traditional data analysis. Data collection, storage, and analysis is a field that is growing very important very quickly: according to International Data Corporation, a market intelligence and analytics firm, worldwide revenues for big data and business analytics will grow from $130.1 billion in 2016 to more than $203 billion in 2020.7


The meteoric rise in importance of this field has smashed existing practices in law firms. Legal practice is one of the most information-heavy industries, with the huge volume of precedent cases, regulations, case analyses, and legal commentaries. It used to be common to see armies of lawyers and paralegals sifting through information as part of due diligence. For the discovery process in United States v. CBS Inc.,8 the studios examined six million documents for more than $2.2 million—much of it going to lawyers and paralegals who worked for months at high hourly rates.9 But that was in 1978. Today, thanks to advances in artificial intelligence, “e-discovery” software can analyse documents in a fraction of the time, for a fraction of the cost. The Californian data analytics firm Blackstone Discovery has software that can analyse 1.5 million documents for less than $100,000, less than a fifth of what it would have cost thirty years ago.

The possibilities of data analytics for lawyers go beyond just search engines; even something as basic as the pricing of legal services has been completely changed. Clients and lawyers now have access to huge databases about prices that firms charge across the jurisdiction. The TyMetrix LegalView service continuously aggregates tens of billions of dollars’ worth of legal invoices, allowing law firms to choose the best position for themselves: an accessible low-cost provider or a high-end premium law firm.10 Another firm, Sky Analytics, offers companies a “Right Rate Advisor” tool, which assesses an external lawyer from many aspects to advise companies whether to accept or reject his fees.11 Over the years, a common complaint levelled at lawyers is the opacity of their fee structures and the resultant uncertainty. With a huge amount of data now available about legal fees, data analysis systems like TyMetrix LegalView and Sky Analytics’ “Right Rate Advisor” can make legal fees more transparent, making legal services more accessible for all.


However, arguably the most important use of data analytics is in the criminal law. Algorithms can now piece together information from a myriad of sources – police reports, arrest statistics, surveillance camera footage, and other information generated by the police – to better understand crime activity in a particular area and, more importantly, where criminals are likely to strike next. Based on this “predictive policing” system, the police can better deploy their resources to areas that are more prone to crime, and can respond faster and more effectively when someone breaks the law. One of the most ambitious predictive policing systems in the world is China’s “Police Cloud” system. Government databases scoop up everything from addresses, to medical history, supermarket membership, and delivery records. This information is linked to each citizen’s unique identification number, and is used by security bureau authorities to look for patterns in an individual’s behaviour. These databases are massive: for instance, police in Jiangsu, China have amassed 780 million data points on its citizens,12 collecting records of citizen’s incomes, navigation data, and their purchases from major e-commerce companies, among other things.13 In Shandong, China, the police can access patient records,14 names and causes of petitioners and political troublemakers,15 and social media usernames. With the Police Cloud, even the most intimate parts of a citizen’s life are open to the government.

The potential for abuse is obvious. In 2015, the Office of the Central Committee of the Chinese Communist Party announced their intention to embrace technology like the Police Cloud to achieve “social stability”. China’s Ministry of Public Security designed the Police Cloud system to surveil seven categories of “focus personnel”, including petitioners, those who “undermine stability”, and people “involved with terrorism”.16 Such vague definitions mean that essentially anyone could be designated a threat and placed under surveillance.

Chinese citizens do not have the right to be notified when placed under surveillance and have no legal avenues for contesting it.17 At present, China has no privacy or data protection law protecting personal data from misuse. The police are under no obligation to obtain a court order to conduct surveillance, or provide any evidence that the people whose data they are collecting from are associated with or involved in criminal activity. There are essentially no effective privacy protections against government surveillance, giving the Chinese police nearly unchecked power.

This is especially alarming if one remembers that predictive policing algorithms are not, and never will be, perfect. Predictive policing systems can only make predictions based on past data, which may not reflect actual risk patterns. Erroneous data will also result in erroneous predictions. A California woman recently won a civil rights lawsuit18 against the San Francisco Police Department after a number-plate reader misidentified hers as a stolen car and she was held at gunpoint by officers, forced to her knees, and detained for 20 minutes.

Especially dangerous is the very real possibility of bias in the input data or the man-made algorithms, causing bias in the predictions. The LAPD has seen a “feedback loop” in its PredPol system sparked by skewed input data.19 A racial bias in the existing crime statistics made the algorithm direct officers to certain neighbourhoods – typically those with many racial minorities – regardless of the true crime rate in that area. Errors and bias in data will inevitably survive in the resulting predictions. Authorities would do well to be cautious in relying too much on predictive policing systems.

That said, predictive policing’s short track record seems promising. Chicago’s 7th District Police reported that shootings in that district dropped 39% from January to July 2017 compared to the same period last year.20 The murder rate also dipped 33% during that period, while the murder rate in the city as a whole rose.21 Predictive policing systems have also beaten human analysts in other real-world trials.22 While the many inherent shortfalls of these systems must be acknowledged, and applications of these systems must take these failures into account, predictive policing is clearly a technology with great potential.


Apart from criticisms of the data analysis systems, data collection on such a scale also raises many legitimate privacy concerns. If GPS data shows Alice was at a hotel at 5 o’clock on Tuesday, and Bob was at the same hotel at the same time, it can be inferred that they might have been together. Further inferences can be drawn from conversation logs between them, if any. Therein lies the core of big data analysis: to reach conclusions that can be drawn through the correlation of many data points, which could not have been drawn from the data points themselves. If these data-sets are from public sources, does the individual have a right to privacy in the new information revealed through analysis? In United States v Maynard,23 the District of Columbia Circuit held that although the appellant’s individual journeys on public roads were public information, since the compilation of data on these journeys would not have been reasonably expected, the appellant’s behaviour patterns revealed through analysis of all of these trips together remain private.

How, then, are the inferences drawn from big data to be kept private? Data-sets are usually anonymized before use to protect the privacy of the individuals that the data was collected from. Anonymization is done through deletion of personally identifiable information (“PII”), or obfuscation thereof (for example, by changing a postcode from 1234567 to 123**). Top government agencies and leading technology companies have embraced anonymization to protect privacy rights: the US Department of Defence has recommended anonymization “whenever practicable”,24 and Google has said that its anonymization techniques can make identification “very unlikely”.25 Prominent legal scholars also share this faith in anonymization,26 and claim that anonymization can make data reidentification “impossible”.27

This faith is misguided. Unfortunately, data reidentification is easier than most would think. If all PII is removed from a data-set, there will be nothing left, since every piece of data is potentially useful in identifying an individual. Therefore, for data, complete privacy means zero utility. For a data-set to have any use at all, some PII must be retained. The widespread faith in anonymization is based on the belief that it is possible to remove enough PII to prevent identification and still retain enough to keep the data-set useful. However, very little PII is actually necessary for deanonymization and thus identification. It has been shown that 87.1% of people in the US can be uniquely identified by their combined five-digit ZIP code, birth date, and sex.28 53% of American citizens are uniquely identified by their city of residence, birth date, and sex.29

Using just three easily obtainable pieces of information, the vast majority of people can be identified using “anonymized” data. Privacy, in the age of big data, is a much more elusive ideal than previously thought. While it would be asking too much for legislators to come up with rules that could prevent any subsequent deanonymization, legislators must at the very least abandon the idea that removing PII is sufficient protection of privacy in today’s security climate. Since this assumption is the foundation to nearly all privacy laws in use today, a paradigm shift is necessary for legislators and industry leaders in this field.


The law is often slow to respond to societal change. Legislatures must mire themselves in debate before promulgating laws far overdue by the time they are passed. The courts, while more flexible, are still reactive rather than proactive. While both bodies stagnate, the total amount of information grows, and private companies and states worldwide are responding to this wealth of information with increasing urgency. In particular, legal practice has been affected: the availability of these huge amounts of data and the analysis has changed case analysis, pricing of legal services, and even the generation of evidence. Police departments worldwide now routinely draw on predictions based on huge data-sets collected from surveillance networks and private companies. Proponents of such policing systems posit this helps the authorities respond faster and more effectively to changing crime trends; critics argue that this breaches the citizen’s right to privacy and gives the authorities untrammelled power to a dystopian degree. Underlying these are the individual’s privacy rights and the unfortunate reality that it is harder to protect these rights than previously thought. Whatever the stance taken, we cannot afford to ignore these issues. A single death is a tragedy but a million deaths a statistic; the countless records of crimes and punishments that form the law have many lessons to teach us, but only if we decide to listen.

[1] Equivalent to 2.5 billion gigabytes, or 2.5 x 1018 bytes.

[2] Ralph Jacobson, “2.5 quintillion bytes of data created every day. How does CPG & Retail manage it?” (24 April 2013), IBM Consumer Products Industry Blog, online: <https://www.ibm.com/blogs/insights-on-business/consumer-products/2-5-quintillion-bytes-of-data-created-every-day-how-does-cpg-retail-manage-it/>.

[3] Ibid.

[4] (No 26 of 2012).

[5] (Cap 50A, 2007 Rev Ed Sing).

[6] From The Oxford English Dictionary, 2nd Ed, sub verbo “big data”.

[7] This is a compound growth rate of 11.7% per annum. See Gil Press, “6 Predictions For The $203 Billion Big Data Analytics Market”, Forbes (20 January 2017), online: <https://www.forbes.com/sites/gilpress/2017/01/20/6-predictions-for-the-203-billion-big-data-analytics-market/#20ece4b72083>

[8] 459 F Supp 832 (CD Cal 1978).

[9] John Markoff, “Armies of Expensive Lawyers, Replaced by Cheaper Software”, The New York Times (4 March 2011), online: <http://www.nytimes.com/2011/03/05/science/05legal.html>.

[10] Joe Dysart, “How lawyers are mining the information mother lode for pricing, practice tips and predictions”,  online: (May 2013) ABA Journal <http://www.abajournal.com/magazine/article/the_dawn_of_big_data/?utm_

[11] Ibid. Sky Analytics has revealed that the Right Rate Advisor refers to, among other factors, the external lawyer’s years of experience, his or her position in the firm, the size of the firm, and the cost of living where the lawyer is based.

[12] What do China’s police collect on citizens in order to predict crime? Everything, online: Quartz <https://qz.com/1133504/to-predict-crime-chinas-tracking-medical-histories-cafe-visits-supermarket-membership-human-rights-watch-warns/>.

[13] Echo Huang, “China: Police ‘Big Data’ Systems Violate Privacy, Target Dissent”, Human Rights Watch (20 November 2017), online: <https://www.hrw.org/news/2017/11/19/china-police-big-data-systems-violate-privacy-target-dissent>.

[14] Including names and illnesses.

[15] Ibid. A tender document from Tianjin boasted that its Police Cloud system could monitor “petitioners who are extremely [persistent]” and “Uyghurs from South Xinjiang”. It could even pinpoint their residences and track their movements on maps.

[16] Ibid.

[17] Ibid.

[18] Green v City and County of San Francisco, 751 F.3d 1039 (2014)

[19] Danielle Ensign et al, “Runaway Feedback Loops in Predictive Policing” (2017) arXiv:1706.09847v2 [cs.CY]

[20] Juliet van Wageren, “Cities Give Predictive Policing a Second Look”, Slate Tech Magazine (12 December 2017), online: <https://statetechmagazine.com/article/2017/12/cities-give-predictive-policing-second-look>.

[21] Ibid.

[22] During a four-month trial in Kent, London, 8.5% of all street crime occurred within and next to the areas the PredPol system designated as high crime areas, predictions from police analysts scored only 5%. An earlier trial in Los Angeles saw the machine score 6% compared with human analysts’ 3%. See “Don’t even think about it”, The Economist (20 July 2013), online: <https://www.economist.com/news/briefing/21582042-it-getting-easier-foresee-wrongdoing-and-spot-likely-wrongdoers-dont-even-think-about-it>.

[23] 615 F.3d 544.

[24] Technology and Privacy Advisory Committee, Safeguarding Privacy in the Fight Against Terrorism (United States of America: Technology and Privacy Advisory Committee, 2004) at 50 (Recommendation 2.2). The document is available at: http://www.cdt.org/security/usapatriot/20040300tapac.pdf.

[25] Chris Soghoian, “Debunking Google's log anonymization propaganda”, CNET (11 September 2008), online: <http://news.cnet.com8301-13 739_3-10038963-46.html>

[26] Ira S. Rubinstein et al, “Data Mining and Internet Profiling: Emerging Regulatory and Technological Approaches” (2008) 75 U Chi L Rev 261, at 266 and 268

[27] Barbara J. Evans, “Congress' New Infrastructural Model of Medical Privacy” (2009) 84 Notre Dame L Rev 585, at 619-20

[28] L. Sweeney, “Simple Demographics Often Identify People Uniquely” (2000) Carnegie Mellon University, Data Privacy Working Paper 3.

[29] Ibid.

The PDF version of this article is available for download here.

Suggestions on the relevance of the Organization's Size to Section 11 of Singapore's Personal Data Protection Act

By Foo Ee Yeong Daniel


Singapore’s Personal Data Protection Act1 [PDPA] has been in effect since four years ago,2 and serves to balance the protection of individuals’ personal data with the ‘need of organisations to collect, use or disclose personal data for purposes that a reasonable person would consider appropriate in the circumstances’.3 Since then, the Personal Data Protection Commission [PDPC], with the help of public consultations,4 has continually revisited and augmented5 the PDPA’s various advisory guidelines.6

The PDPA stipulates various obligations, which organisations should fulfil based on ‘what a reasonable person would consider appropriate in the circumstances’, as per section 11(1) of the Act (the “reasonableness test”). This standard of reasonableness underpins the standard of compliance for all obligations under the PDPA,7 and the Advisory Guidelines on Key Concepts in the PDPA8 [Guidelines] clarifies that this applies to all private organisations9 as defined in section 2, regardless of their size.

This article aims to explore the relevance of an organisation’s size to the PDPA’s reasonableness test, and submits that the former should be considered as a factor in applying the latter. This may be done, for instance, by providing for it in the Guidelines.


Currently, the size of the organisation appears to be contemplated only by the Protection Obligation – when determining whether reasonable security arrangements have been made to prevent unauthorised handling of personal data under section 24 of the PDPA. This is seen from the Guidelines, which provide only general guidance for compliance10 and mention the size of the organisation only once: as a factor in risk assessment exercises determining whether information security arrangements are adequate.11 Otherwise, the PDPA legislation and jurisprudence do not feature the size of the organisation in applying the reasonableness test for any other obligation. There does not appear to be any debate on this issue; one can only guess that the drafters of the PDPA believed that fulfilling these other obligations was more important than the strain of compliance on organisations and/or that the obligations were generally undemanding for organisations that already had strong data protection practices. In any case, the PDPA prevents organisations from invoking their small size to unjustifiably exempt themselves from obligations to protect personal data.

Instead, the reasonableness of measures appears to turn on the impact on the individual whose personal data is mishandled and the compliance measures themselves. For example, the Accuracy Obligation under section 23 considers, inter alia, the nature of the personal data,12 as well as the impact on the relevant individual should the data be inaccurate.13 Another example is the Notification Obligation under section 20, which considers the ‘circumstances and manner in which [the organisation] will be collecting the personal data’,14 the ‘frequency at which the personal data will be collected’15 and ‘the channel through which the notification is provided’.16 The size of the organisation does not appear to feature in the reasonableness test for any of the obligations under the PDPA, except for in the Protection Obligation.


As a result of the above, the reasonableness test arguably fails to take into account the resource-scarce reality of many small organisations when determining whether they have discharged their obligations to a ‘reasonable’ standard under the PDPA. One example is where an organisation transfers personal data to its parent company overseas, and has to fulfil its Transfer Limitation Obligation under section 26 of the PDPA. The Guidelines suggest that the organisation reviews the corporate rules binding both organisations and assesses that they comply with these regulations, as well as that the data protection is ‘comparable to the standard under the PDPA’.17 This envisages studying rules, designing and executing appropriate transfers, as well as deciding whether corporate practices sufficiently comply with legislation – all difficult processes that require a certain amount of manpower or at least expertise that small organisations will not be as privy to as large ones. Except for in the Protection Obligation, the PDPA’s current reasonableness test essentially demands the same standard of compliance from the sole proprietor as that from the large, multinational company. This raises issues of resource inequality and disadvantage to small organisations, for which sustainability is already a challenge without the PDPA.

Considering the organisation’s size when applying the reasonableness test would better accord with the plain meaning of ‘reasonableness’. It appears unreasonable, in the barest and most layman sense of the word, to expect small organisations to comply with the PDPA as rigorously as large organisations. Then-President of the Singapore Chinese Chamber of Commerce and Industry, Mr Teo Siong Seng, emphasised during the Second Reading of the Personal Data Protection Bill that small organisations would struggle more with manpower, time-related and even consultancy costs of compliance with the PDPA.18 SMEs have since reportedly had to grapple with ‘overburdened staff’19 and five-figure costs on ‘new procedures, staff training and the upgrading of technology]’.20 In particular, the obligation to ‘develop and implement policies and practices that are necessary’ to comply with the PDPA, as per section 12(a) of the Act, is manifestly more difficult for small organisations than it is for large ones. Taking into account an organisation’s size would achieve better approximations of what a ‘reasonable person would consider appropriate in the circumstances’. This would in turn produce more practical benefits: guiding the PDPC to achieve fairer adjudicative outcomes – ensuring that small organisations are not penalised for failing to take compliance measures beyond their means.

Further, having regard for the size of the organisation would better achieve the PDPA’s purpose of mitigating compliance costs.21 Organisations should save costs when implementing essential PDPA-compliant processes, as doing so guards against actionable, personal data breaches ‘under other statutes, at common law and equity’.22 This helps organisations save costs on litigation and compensation, which would be greater than the costs incurred for compliance with the PDPA. However, as the PDPA’s reasonableness test now apparently does not accommodate the inherent differences between small and large organisations, small organisations may find themselves tending toward the safest practices or ‘best solution[s]’ adopted by large organisations, which may be too costly for them.23 Recognising that the size of the organisation should affect what is considered ‘reasonable’ compliance would give a green light to small organisations and their consultants (if they can afford any) to exercise latitude in adopting more cost-efficient practices that would still comply with the PDPA.24

Considering the organisation’s size would also better achieve the PDPA’s purpose of enhancing Singapore’s business competitiveness.25 Holding small organisations to the same standard of ‘reasonableness’ as large organisations in complying with the PDPA has deleterious effects on the former’s operations.26 This is because compliance with the PDPA requires a large amount of time, cost and effort that could otherwise be invested productively into the organisation’s operations.27 Such resource-demanding measures include studying the PDPA, appointing a Personal Data Officer,28 developing policies and practices for compliance29 that must be then communicated to staff,30 as well as training staff to receive and respond to PDPA-related inquiries and complaints.31 This has arguably even worse consequences for small social service organisations, which already struggle to make the most of their resources to perform their charitable works. Imposing the reasonableness test for compliance – without considering their sizes – risks impeding the good work of these organisations and generally inhibiting the progress of Singapore’s social service sector – a result that is normatively undesirable. A reasonableness test that accounts for the organisation’s size would encourage small organisations to consider practices that are less operationally disruptive than those that large organisations adopt but would still comply with the PDPA. Such would not only facilitate the Act’s aim of business productivity and competitiveness, but also the socially desirable aims of various social service organisations.


The dangers of considering an organisation’s size in applying the reasonableness test may be observed from the effects of Australia’s Privacy Act 198832 [APA], which makes exemptions for small businesses.33 This ‘small business exemption’ has been heavily criticised, and the Australian Law Reform Commission even recommended its repeal in 2008.34 In particular, it has been argued that organisational size is unrelated to the risk of personal data breach; such depends instead on the nature of the data, its handling and the organisation’s operations.35 There have been concerns that the APA may be abused by small organisations, which are given a statutory backdoor to misuse personal data in the name of cost-effectiveness.36

Including the organisation’s size as only a factor in the PDPA’s reasonableness test would be an appropriately moderate approach that mitigates the risk of completely exempting small, rogue organisations. In fact, as suggested, the organisation’s size could be mentioned as a factor only in the Guidelines. Since the Guidelines are ‘advisory’ and ‘do not constitute legal advice’,37 this would mitigate the risk of giving small organisations carte blanche to breach the PDPA – holding small organisations to baseline standards of compliance. The non-conclusive status of a ‘factor’, as well as the non-binding nature of the Guidelines, also collectively preserve the PDPC’s ability to find PDPA breaches regardless of the organisation’s size. The only difference would be that the PDPC should be persuaded to weigh the small size of the organisation as one of many factors in deciding whether there is a breach of the Act.


In sum, an organisation’s size should be considered as a factor in the PDPA’s reasonableness test as such better accords with the plain meaning of ‘reasonableness’, as well as better achieves the Act’s purposes of enhancing Singapore’s business competitiveness while managing compliance costs.

Having explored the legal and normative justifications of incorporating the organisation’s size as a factor in the PDPA’s reasonableness test, this article notes that defining ‘size’ has proven and can be expected to be tricky. The Australian Privacy Act’s definition of a ‘small business’ may be used as a case study. It sets out what would and would not qualify for the exemption, and has two significant features: first, it pegs ‘size’ primarily to the organisation’s annual turnover.38 Second, it adopts a binary view of what would be ‘small’ and not. Suggestions have been made to raise the Privacy Act’s turnover threshold, to account for inflation.39 There have also been suggestions to base the definition instead on specific levels of risk 40 or simply the number of employees in the organisation.41 Each of these has attracted its criticisms.

Thus, careful thought should be given as to what definition of ‘size’ would be a suitable factor in the PDPA’s ‘reasonableness’ test, considering the Act’s aim of promoting business competitiveness and data protection while moderating compliance costs. These, as well as other matters related to how the size of the organisation may or should affect its compliance obligations, should also be further considered.

[1] Personal Data Protection Act (No. 26 of 2012).

[2] According to the Personal Data Protection Commission of Singapore, “Legislation and Guidelines”, online: <https://www.pdpc.gov.sg/legislation-and-guidelines>, ‘[t]he PDPA took effect in phases starting with the provisions relating to the formation of the PDPA on 2 January 2013. Provisions relating to the DNC Registry came into effect on 2 January 2014 and the main data protection rules on 2 July 2014.’

[3] Supra note 1 at s 3.

[4] A useful repository of these consultation papers may be found at the Personal Data Protection Commission of Singapore, “Public Consultations”, online: < https://www.pdpc.gov.sg/legislation-and-guidelines/public-consultations>.

[5] Singapore Parliamentary Debates, Official Report, vol 93 (10 March 2015) (Assoc Prof Dr Yaacob Ibrahim).

[6] Supra note 1 at s 49(1) states that the ‘Commission may, from time to time, issue written advisory guidelines indicating the manner in which the Commission will interpret the provisions of this Act’.

[7] These may be broadly labelled as the Consent Obligation, Purpose Limitation Obligation, Notification Obligation, Access and Correction Obligation, Accuracy Obligation, Protection Obligation, Retention Limitation Obligation, Transfer Limitation Obligation and Openness Obligation. This article will look at only a few of these obligations, mostly for illustrative purposes.

[8] Personal Data Protection Commission of Singapore, “Advisory Guidelines on Key Concepts in the PDPA” (revised 27 July 2017).

[9] Ibid at 6.3.

[10] Personal Data Protection Commission of Singapore, “Introduction to the Guidelines” at para 3.3.

[11] Supra note 8 at 17.4(a).

[12] Ibid at 16.4(a).

[13] Ibid at 16.4(e).

[14] Ibid at 14.10(a).

[15] Ibid at 14.10(c).

[16] Ibid at 14.10(d).

[17] Ibid at 19.4.

[18] Parliamentary Debates Singapore: Official Report, vol 89 (15 October 2012) (Nominated Member, Mr Teo Siong Seng). Mr Teo spoke as then-President of the Singapore Chinese Chamber of Commerce and Industry, ‘representing 4,000 corporate members and 145 trade associations from a great diversity of trades, industries and service providers’.

[19] The Straits Times, “Privacy Act sows confusion”, online: < http://www.todayonline.com/commentary/privacy-act-sows-confusion>.

[20] The Straits Times, “Early childhood educator simplifies personal data protection requirements”, online: <http://www.straitstimes.com/business/companies-markets/early-childhood-educator-simplifies-personal-data-protection-requirements>.

[21] Singapore Parliamentary Debates, Official Report, vol 89 (15 October 2012) (Minister for Communications and Information, Associate Professor Dr Yaacob Ibrahim).

[22] Hannah YeeFen Lim, Data Protection in the Practical Context: Strategies and Techniques (Singapore: Academy Publishing, 2017) at 1.2.

[23] Supra note 19.

[24] It is interesting to note that organisations are already advised to consider their size in deciding appropriate audit processes, as per Alat Sheela, Role Of Audit In Your Organisation’s Personal Data Protection Act 2012 Compliance Programme, Personal Data Protection Digest (Singapore: Academy Publishing, 2017) at 10. It is submitted that this should be the case for all forms of compliance practices.

[25] Supra note 22.

[26] This was recently alluded to in Parliamentary Debates Singapore: Official Report, vol 94 (6 March 2017) (Member of Parliament, Mr Saktiandi Supaat): the need to protect personal data has made it ‘more difficult’ for businesses to ‘use data innovatively and optimize business opportunities’.

[27] Supra note 19.

[28] This is an extension of the PDPA’s obligations; supra note 1 at s 11(3).

[29] Ibid at s 12(a).

[30] Ibid at s 12(c).

[31] Ibid at s 12(b).

[32] Privacy Act 1988 (Cth).

[33] Ibid at s 6C.

[34] See Recommendation 39-1 in Australian Law Reform Commission, “Australian Privacy Law and Practice Report 108” at p 53.

[35] Ibid at 39.26.

[36] A more detailed analysis of the Privacy Act 1988’s ‘small business exemption’ may be found in Supra note 35.

[37] Supra note 10 at 3.1.

[38] Supra note 32 at s 6D.

[39] Supra note 34 at 39.124.

[40] As determined by the type of data and number of individuals about whom data is held; Ibid at 39.126.

[41] Ibid at 39.129.

The PDF version of this article is available for download here.

New Sentencing Principles for Rape – An Analysis of Ng Kean Meng Terence v Public Prosecutor

By Benjamin Low


What is the most appropriate sentence that should be meted out onto an accused person who has been found guilty of the offence of rape?

The apparent simplicity of this question belies powerful, and oftentimes conflicting, sentencing policy considerations and more abstract questions of ethical principles and sentencing theory. The recent case of Ng Kean Meng Terence v Public Prosecutor1 [Ng Kean Meng Terence] is the Singapore Court of Appeal’s latest attempt at creating a new, comprehensive sentencing framework for rape in order to address the above quandaries. This article seeks to analyse the state of the law prior to Ng Kean Meng Terence and discusses the implications and effects of the case on the criminal law.


Prior to Ng Kean Meng Terence, the law on sentencing for rape offences was encapsulated in Public Prosecutor v NF2 [NF]. In NF, V K Rajah J (as he then was), formulated a new sentencing framework (the “NF Framework”) that was meant to replace the earlier sentencing guidelines in the earlier case of Chia Kim Heng Frederick v Public Prosecutor3 [Frederick Chia].

Under the NF Framework, rape cases were divided into four categories, each with a minimum sentence of imprisonment and caning of differing degree.4 The NF Framework may be briefly summarised as follows:

  1. Category 1 – Cases with no aggravating and/or mitigating factors with a starting sentence of ten years imprisonment and not less than six strokes of caning.
  2. Category 2 – Cases with any one of the seven aggravating factors listed in NF, such as where rape that is committed by two or more persons acting in concert; or where the offender is in a position of responsibility towards the victim.5 For such instances, the starting sentence would have been fifteen years imprisonment and twelve strokes of the cane.
  3. Category 3 – Cases where the victim is raped on multiple occasions, or where rape is perpetrated on multiple victims. This category shares the same sentencing benchmark as a Category 2 case. Rajah J justified this on the basis that the Prosecution would normally prefer multiple charges against the offender, thereby leading the courts to order that two or more of the sentences imposed run consecutively, thereby resulting in a higher sentence meted out.6
  4. Category 4 – Cases where the offender has “manifested perverted or psychopathic tendencies or gross personality disorder, and where he is likely, if at large, to remain a danger to women for an indefinite time”.7 In such cases, the maximum sentence of twenty years imprisonment and twenty-four strokes of the cane would be handed out.

Ideally, the NF Framework operates by having the Court first determine the category under which the particular rape offence should be placed, followed by the Court then adjusting the benchmark sentence upwards or downwards to take into account the additional aggravating and mitigating factors disclosed on the facts.8

The NF Framework was first considered and approved of by the Court of Appeal in the context of rape simpliciter in Public Prosecutor v Mohammed Liton Mohammed Syeed Mallik.9 It was subsequently applied by the Singapore courts to a wide range of cases involving rape committed in various permutations and factual situations.10


Notwithstanding the durability and versatility of the NF Framework over the past eleven years, the Court of Appeal in Ng Kean Meng Terence noted that problems remained with the NF Framework that, in its opinion, necessitated a revision of the sentencing law on rape. These were, firstly, that the categories laid down in NF were themselves not properly defined – to the extent that some of the categories served no clear purpose.11 Category 3, for example, was defined in NF as encompassing cases involving the repeated rape of the same victim or rapes committed by a single offender on multiple victims. However, given that the Criminal Procedure Code12 requires that (and tying in with longstanding prosecutorial practice) separate and multiple charges be pressed against an accused person based on the facts disclosed13 and separate sentences imposed for each charge which the accused person is convicted of,14 this appeared to remove the need for a separate Category 3 to deal with multiple offending under the NF Framework.

A second, more pressing problem, was the lack of conceptual coherence with which the Category 2 aggravating factors were included inside Category 2. While it is not possible to list out all the Category 2 aggravating factors here, the Court of Appeal noted with disquiet that “[T]here does not appear to be any conceptual unity or discernible unifying theme”15 underlying the inclusion of these factors and the exclusion of others. While Rajah J no doubt intended that the list of Category 2 aggravating factors themselves be complemented with the additional aggravating factors and/or mitigating factors of each specific case, there is considerable weight in the Court’s criticisms with regards to the admittedly unclear criteria on which the Category 2 aggravating factors were based on.

The final nail in the coffin was the Court of Appeal’s finding that the NF Framework did not properly account for the statutory aggravating factors already present in the Penal Code. Section 375(3) of the Penal Code16 provides for a minimum sentence of 8 years imprisonment and 12 strokes of the cane if an offender either (a) voluntarily causes hurt to a victim or any other person in order to commit or facilitate the commission of rape; (b) puts the victim in fear of death or hurt to herself or any other person in order to commit or facilitate the commission of rape; or (c) rapes a victim under 14 by having sexual intercourse with her without her consent. None of these factors are included in the Category 2 aggravating factors within the NF Framework. In the Court’s opinion, there was a need to clarify the relationship between the two groups of aggravating factors.


Having opted to do away with the NF Framework, the Court in Ng Kean Meng Terence embarked on the difficult task of developing a new sentencing framework (the “Ng Kean Meng Terence Framework”) for the offence of rape. After considering past approaches by the courts in the preceding case law, the Court of Appeal rejected these approaches as unsuitable and opted to adopt the framework developed in the New Zealand case of R v Taueki.17

Generally speaking, the Ng Kean Meng Terence Framework operates in two steps:18

  1. Firstly, the court should identify which sentencing band the offence in question falls within by taking into account the offence-specific factors in the case. Once the sentencing band has been identified, the Court should then determine where within the aforementioned band the present offence falls in order to derive a sentence for a starting point.
  2. Secondly, the court should then consider the aggravating and mitigating factors which are personal to the offender to calibrate the appropriate sentence for the offender.

Step 1. Classification of the Offence

The Court of Appeal stressed that only “offence-specific” factors which relate to the circumstances of the offence such as the particular harm caused, the manner in which the offence was committed, or the specific role played by the offender in the commission of a offence by a group.19 This proposition is a common-sensical one that recognises that different factual permutations of rape should attract different penal consequences.

A non-exhaustive list of offence-specific factors, drawing on existing case law, was proposed by the Court:20

  • Group Rape: Offences committed by groups of persons.
  • Abuse of position and breach of trust: Where the offender is in a position of responsibility towards the victim or in whom the victim has placed her trust by virtue of his office of employment.
  • Premeditation: Where the offender engages in some form of planning or exhibits premeditation in committing the act (e.g. through the use of drugs or engaging in predatory behaviour).
  • Violence: Actual or threatened use of violence.
  • Vulnerable victim: Whether the victim was especially vulnerable by reason of age, physical frailty, mental impairment or disorder, and/or learning disability.
  • Forcible rape of a victim below the age of fourteen.
  • Hate crime: Where the offence was committed as a hate crime, such as an expression of racial and/or religious prejudice, or in other situations where the victim has been specifically targeted by reason of her membership of a vulnerable minority group.
  • Severe harm: Severe harm inflicted on the victim such as pregnancy, transmission of a serious disease or a psychiatric illness.
  • Deliberate infliction of special trauma.

The Court of Appeal also cautioned against taking into account the two factors of (i) forgiveness by the victim towards the offender and (ii) consent given by a victim under fourteen on the basis that such factors were usually irrelevant.21 Rejecting the second factor in particular, ties in with the express wording of s 375 of the Penal Code in criminalising sex with a person under the age of fourteen.

Having regard to the offence-specific factors, the Court must then place the offence within an appropriate sentencing band.22 For ease of reference, the three sentencing bands, each with different sentencing outcomes, are laid out below:

Table_Sentenncing Bands.PNG

Once the appropriate sentencing band has been identified, the Court would then identify where precisely along the range prescribed for the band a particular sentence falls.24

Step 2. Calibrating the Sentence

Once the starting-point sentence has been identified, the Court must then consider the “offender-specific” factors of the case – that is to say, those factors which relate to the personal circumstances of the offender himself,25 rather than relating to the manner and mode of the offending. In making this distinction, it seems that the Court of Appeal was clearly seeking to prevent any double-counting of factors that could lead to an imposition of more severe sentences than would usually be the case.

The possible offender-specific aggravating and mitigating factors recognised in Ng Kean Meng Terence are as follows:26

Table_Offender-specific factors.PNG

One of the more controversial aspects of Ng Kean Meng Terence was whether pleas of guilt could constitute a mitigating factor. While this contentious point will be discussed in further detail below, it will suffice to say that the Court in Ng Kean Meng Terence was prepared to treat pleas of guilt as one of the many offender-specific mitigating factors.27


From the outset, it should be noted that the Ng Kean Meng Terence Framework does not constitute a fundamental or radical revision of the sentencing law for rape. What the Ng Kean Meng Terence Framework does is to improve the present sentencing framework by better delineating the different factors that courts usually consider when determining the appropriate sentence to be meted out onto the offender.

Thus the distinction between “offence-specific” and “offender-specific” factors removes the problem of double-counting while ensuring that all relevant factors in each case are taken into consideration by the courts.28 The offence-specific factors themselves have also been modified to circumvent the conceptual difficulties of the NF Framework while addressing the third problem raised as to the statutory-aggravating factors in the Penal Code.29

A second noteworthy point is the use of sentencing bands which injects greater clarity in determining the starting-point sentence. Rather than having to work from scratch at a fixed benchmark sentence under the NF Framework, sentencing courts can now better identify which Band, and thus which part of the Band, the case in question belongs to.

It bears mentioning that since the decision of Ng Kean Meng Terence, there have been some additional developments in the local case law with regards to the Ng Kean Meng Terence Framework. While it is not possible to consider these developments in great detail, they warrant a mention in this article for completeness sake. In the later decision of Pram Nair v Public Prosecutor,30 the Court of Appeal recognised that the intoxication of a victim could constitute an additional offence-specific aggravating factor though it cautioned that this factor would not necessarily manifest itself in all cases involving an intoxicated victim.31 Much would depend on the degree of intoxication in each particular case. More importantly, the Ng Kean Meng Terence Framework was transposed onto the offence of sexual assault by penetration, with three sentencing bands imposed with varying sentencing ranges.32 As a preliminary note, these developments ought to be welcomed as promoting unity in the courts’ approaches towards both types of sexual offences.

However, notwithstanding the merits of the Ng Kean Meng Terence Framework, it is submitted that the shift in the Court of Appeal’s attitude towards pleas of guilt could carry possibly adverse implications in the sentencing law towards criminal offences. Prior to Ng Kean Meng Terence, it was settled law that a plea of guilt would not entitle an offender to a sentencing discount unless it was made out of genuine remorse.33 However, the Court of Appeal has since departed from that position by recognising in Chang Kar Meng v Public Prosecutor34 [Chang Kar Meng] that a plea of guilt could constitute a mitigating factor for utilitarian reasons (i.e. it saves victims the horror of having to recount the ordeal and it saves the resources of the State if the case were to go to trial) besides the existing remorse-based justification. The Court galvanized this proposition in Ng Kean Meng Terence by citing Chang Kar Meng with approval.35

While the utilitarian considerations admittedly cannot be discounted, the Court of Appeal’s pronouncement in Ng Kean Meng Terence has the effect of suggesting that a plea of guilt will almost certainly entitle an offender to a sentencing discount as long as he pleads guilty to a charge of rape, even if the guilty plea was not motivated by genuine remorse or contrition. Though the Court later sought to clarify that guilty pleas would only be treated as one of the many mitigating factors, it is not difficult to foresee that rape offenders would be more inclined to try their luck and enter pleas of guilt in the hopes of obtaining a lighter sentence. In a Band 1 case of rape with no further aggravating factors, a plea of guilt could potentially entitle an offender to a sentence beneath the lowest threshold in Band 1. In such situations, and where the offender does not seem to have displayed genuine remorse, can it truly be said that the public interest has been effectively safeguarded or that the overriding considerations of rehabilitation and/or deterrence have been achieved?

This leads us to a more fundamental problem with the Ng Kean Meng Terence Framework in that it does not provide any real change in the sentencing law for rape. While this was no doubt the express intention of the Court of Appeal itself,36 one cannot help but feel that a golden opportunity was missed in reviewing the present sentencing law on rape and then determining whether or not the prescribed sentencing ranges ought to be revised upwards or downwards. Indeed, the lowest sentence of ten years imprisonment and six strokes of caning in Band 1 was clearly based on the earlier decision in Frederick Chia37 but this writer submits that the sentencing ranges for the various sentencing bands should be revised upwards. A stricter sentencing regime would better accord with the longstanding deterrent aspect of the sentencing jurisprudence of the Singapore courts,38 especially given a disturbing rise in sexual crimes over the past few years.39


The process of determining and laying down suitable and comprehensive sentencing guidelines for various offences is a continuous and challenging effort that often requires the courts to continuously take into account broader societal and legal changes in an effort to ensure that such guidelines remain relevant and up-to-date while ensuring that the public interest is safeguarded. Since the NF Framework, the criminal law in Singapore has evolved immensely, and the Courts must continue to refine and redevelop the existing case law to account for the new legal climate.

In light of the intrinsic seriousness of the offence of rape in particular, and the difficulties that courts sometimes encounter in threshing out the most appropriate sentence, the Court of Appeal has boldly decided to revise the sentencing law for rape by devising a new sentencing framework that meets these requirements. Although it remains to be seen whether the Ng Khean Meng Terence framework is sufficiently capable of withstanding future challenges in the form of hard cases, there is much to commend the Court for its efforts in tackling what has been a most complex area of criminal law.

[1] [2017] 2 SLR 449.

[2] [2006] 4 SLR(R) 849.

[3] [1992] 1 SLR(R) 63.

[4] Supra note 2 at [20]–[21].

[5] Ibid.

[6] Ibid, at [37].

[7] Ibid, at [21]; see also R v Keith Billam (1986) 8 Cr App R (S) 48 at pp 50–51.

[8] Ibid, at [23]; see also Ng Kean Meng Terence at [10].

[9] [2008] 1 SLR(R) 601.

[10] See PP v UI [2008] 4 SLR(R) 500 and PP v AOM [2011] 2 SLR 1057.

[11] Supra note 1 at [13]–[15].

[12] (Cap 68, 2012 Rev Ed).

[13] Ibid, at s 132.

[14] Ibid, at s 306(1).

[15] Supra note 1 at [17].

[16] (Cap 224, 2008 Rev Ed).

[17] [2005] 3 NZLR 372.

[18] Supra note 1 at [39].

[19] Ibid, at [43].

[20] Ibid, at [44].

[21] Ibid, at [45].

[22] Supra note 1 at [47].

[23] See PP v Bala Kuppusamy [2009] SGHC 97 at [28]–[29].

[24] Supra note 1 at [61].

[25] Ibid, at [62].

[26] Ibid, at [64]–[65].

[27] Ibid, at [71].

[28] Ibid, at [73].

[29] Supra note 16 at s 375(3).

[30] [2017] 2 SLR 1015.

[31]Ibid, at [126]–[132].

[32] Ibid, at [158]–[159]; although, strangely, the Court held that the respective sentencing ranges for sexual assault by penetration were to be lower than that for rape.

[33] Supra note 1 at [67]; see also Rajah J’s holding in Angliss Singapore Pte Ltd v PP [2006] 4 SLR(R) 653 at [56] and [77].

[34] [2017] 2 SLR 68 at [47].

[35] Supra note 1 at [68]–[69]; see also Chia Kim Heng Frederick v PP [1992] 1 SLR(R) 63 at [20].

[36] Ibid, at [74].

[37] Supra note 3 at [20].

[38] See PP v Law Aik Meng [2007] 2 SLR(R) 814 at [18]–[19].

[39] Seow Bei Yi, “Rise in sex crimes over past 5 years: State Courts”, The Straits Times (3 April 2017), online: <http://www.straitstimes.com/singapore/courts-crime/rise-in-sex-crimes-over-past-5-years-state-courts> ; see also “Violence: Rape Victims” (24 February 2017), online: Ministry of Social and Family Development <https://www.msf.gov.sg/research-and-data/Research-and-Statistics/Pages/Violence-Rape-Victims.aspx> .

A PDF version of this entry is available for download here.

Economic Analaysis of Law with Indian Characteristics: Shiva Shakti Sugars Ltd. v Shree Renuka Sugar Ltd [2017]

By Darren Teoh


In the landmark case of Shivashakti Sugars Ltd. v Shree Renuka Sugar Ltd.1 [Shiva Shakti], the Supreme Court of India opined that the “economic analysis of law” ought to be broadly received into India, and serve as a guide to statutory interpretation. This article primarily focuses on how the Court has defined the “economic analysis of law” as it ought to be applied within India, since it differs radically from the original (American) conception of such an analysis. A brief critique of the means-ends rationality2 of the Court’s construction of the Indian “economic analysis of law” is laid out towards the end of the article.


A. Facts and Procedural History

The plaintiff, Shri Renuka Sugar Mills Ltd, brought a suit in the High Court of Karnataka against the defendant, Shiva Shakthi Sugar Ltd. The plaintiff argued that since the place where the defendant had set up its sugarcane-crushing factory was within 15 kilometres of its own, “as per the provisions of clause 6A of Sugarcane (Control) Amendment Order, 2006, no permission could have been given to the appellant to start its factory”. The High Court ruled in favour of the plaintiff and ordered that the defendant cease its sugar-cane crushing operations. The Court held that since the factory operated by the plaintiff was an “existing sugar factory” when the defendant had applied to set up a factory by means of an Industrial Entrepreneurs' Memorandum (“IEM”), Clause 6A’s prohibition was attracted.4 Further, the two extensions to the defendant’s IEM, made in August and December 2010 respectively, were found to be invalid since the defendant had not taken “effective steps” to justify the extensions, as per the requirements of Clause 6C. The defendant appealed to the Supreme Court of India.

B. Decision of the Supreme Court

The Supreme Court allowed the defendant’s appeal. The Court identified seven legal issues raised by the appeal, alphabetically enumerated (a) to (g),5 but only addressed two of the seven aforementioned issues, (b) and (g). It opined that the case would be satisfactorily disposed upon answering (b) and (g).6

Issue (b) pertained to whether the respondent’s factory could be brought within the ambit of Clause 6A’s prohibition as an “existing sugar factory”. This relied on a construction of Clause 6A in relation to Explanation 1 of the Sugarcane (Control) Amendment Order (“SCAO”), and an assumption that Clause 6A applies to an IEM holder retrospectively. The Supreme Court determined that the respondent’s factory was not an “existing sugar factory”. It held that the High Court had erred in recording that the requirement in Explanation 1 of Clause 6A could be met if sugarcane-crushing had occurred in any one of the five seasons. Instead, the SCAO required factories to crush sugar for five consecutive years before they qualified as “existing sugar factories”.7

Even if the High Court had not erred, there was an antecedent Issue (g) – whether the appellant’s factory should be allowed to continue its business in the interests of justice. The court deliberated on Issue (g) in light of the subsequent establishment and operation of the appellant’s factory, but declined to consider whether “effective steps” were taken as required by Clauses 6A, 6B and 6C of the Sugarcane Control order. Given that the appellants had spent 3 billion rupees (over 62 million Singapore dollars) on setting up the factory, provided direct employment to roughly 700 workers etc., the Supreme Court held that the appellant’s factory should be permitted to continue operating even if the lower Court had not erred and the respondent was operating an “existing sugar factory”.8

The Supreme Court’s finding on Issue (g) was grounded on a liberal approach to statutory interpretation. Although the general preference is to give effect to the plain meaning of a statutory provision, the Court retained discretion to construe the provision flexibly, where there was ambiguity as to the meaning of the words in the statute, or if the legislature had delegated discretion. If so, the provision would be interpreted in a way that would promote an outcome which was beneficial from an economic perspective.9

In exercising its discretion in constructing the provision, the Supreme Court relied on three separate, but inter-connected justifications. It first remarked that it had the duty to support the liberalisation of India’s economy and promote economic growth alongside the other branches of Government.10 Additionally, the Supreme Court opined that the “economic analysis of law” was a valid jurisprudential approach to interpretation, and noted its far-reaching influence in the United States. Even if there was a “technical violation”, the Supreme Court asserted that it had both the power and responsibility under article 142 of the Indian Constitution to bypass the strict requirements of the statute, as long as it was in the economic interests of the nation.11


A. Comparing Indian and American Understandings of Economic Analysis of Law

The Supreme Court observed that American courts have utilised economic analysis in multiple areas of law.12 However, American sources, notably those of Richard Posner, were only used in support of the general proposition that the potential economic effects of rulings ought to be taken into account during a court’s deliberations. Indeed, the Supreme Court categorically stated that it was not discussing the acceptability of the normative and positive theories of “economic analysis of law”.13 In referring to the terms “positive” and “normative”, the Court clearly distinguished its position from that in American nomenclature.

What then are the differences between the Indian and American understandings of economic analysis of law? Two observations can be gleamed from the judgement.

First, the Indian approach to economic analysis of law extends readily to statutory interpretation, whereas such had been mostly limited to common law decision-making in the United States. American Courts have only ventilated their adoption of economic analysis where statutes concerned were open-textured, and if it was within reasonable contemplation that the legislature had delegated law-making power to the courts. These statutes are known as “common-law statutes”, as they implicitly permit judges to develop a common-law around the statute or its particular provisions.14 An appropriate example will help bring the difference into sharp focus – Section 1 of the United States’ Sherman Act regulating anti-competitive behaviour states that any “contract, combination . . . or conspiracy . . . in restraint of trade”15 is unlawful. Such a statute would require economic analysis, because plain reading offers no guidance as to whether a specific business practice violated section 1. It would be necessary for the Court to provide further qualitative judgement as to how the section ought to apply. On the other hand, Clause 6A of the Sugarcane (Control) Amendment Order (2006) would not require economic analysis under the American approach. Where it states “…no new sugar factory shall be set up within the radius of 15 km of any existing sugar factory or another new sugar factory in a State or two or more States”16; it is evident that not much interpretation is required to understand the clause’s substantive meaning.

The second difference between American and Indian “economic analysis of law” is that in the former, the methodology of neo-classical microeconomic theory is applied in analysing and developing legal rules, and the ends sought are welfare maximization and economic efficiency. In the latter, the individual’s strict legal rights have to be balanced against the need to guarantee distributive justice through the promotion of the nation’s economic development. This is evident from the Supreme Court’s consideration of a broad range of factors including job opportunities created, the executive branch’s approval of the business venture and the large costs already incurred in coming to its decision.17 Further support for the author’s framing of the Indian approach is that the Supreme Court also remarked that “Even in those cases where economic interest competes with the rights of other persons, [the] need is to strike a balance between the two competing interests and have a balanced approach”.18

B. A Critique of the Decision’s Means-Ends Rationality

The aversion of Commonwealth and American courts towards interfering with economic policy-making by the elected branches of Government19 might not make sense in India from a legal (as opposed to ethical or political) perspective. This is because the preamble of India’s constitution describes the country as a “sovereign, socialist secular democratic republic”. Holmes’ sweeping dictum in New York v Lochner, that a “constitution is not meant to embody a particular economic theory”, does not necessarily apply in India.20

In any case, how the preamble ought to influence judicial decision-making is a matter of constitutional interpretation, which lies beyond the scope of this article. The author is content with answering a more modest inquiry, namely, “how well does the Indian approach to economic analysis of law as articulated in Shiva Shakti support the judiciary’s goal of promoting the growth and liberalisation of the economy?” In brief, because the Indian approach to economic analysis incorporates distributive justice concerns, there will necessarily be a great tension in the law when there is no confluence between the nation’s economic interest and distributive justice. In Shiva Shakti, no issue in this respect arose. Allowing the appeal maximized allocative and productive efficiency since more than enough sugarcanes were available for crushing in the area. Distributive justice was also preserved because the presumably low-income factory workers retained their jobs. But it is not difficult to imagine situations under which the Court will find itself hard pressed to sacrifice one for the other.21

Moreover, there is a real risk of a great flood of litigation swamping the courts. This is because an absence of a finer balancing test to guide lower courts in applying the “economic analysis of law” and the assertion that courts can rule contrary to the express dictates of statute will make it difficult to predict whether one’s actions are lawful in a particular factual circumstance. The author foresees that there will be an increase in litigation brought to the courts to clarify Shiva Shakti. As the Indian court system is already over-burdened by a backlog of cases,22 it is hard to see how the nation’s economic growth and liberalisation will be better promoted by a crudely fashioned approach to “economic analysis of law”. India’s Chief Justice T.S. Thakur’s observation in 2016 that foreign investors and manufacturers are rightly concerned about the ability of the Indian judicial system to deal with disputes efficiently is apposite here.23


The decision in Shiva Shakti accords with the Indian Supreme Court’s activist character, and reflects the court’s desire to promote India’s economic development. However, it is doubtful that the Court’s assumption of a larger role in economic policy-making would be an effective means of achieving that goal. This is because judges are largely appointed from legal practice or academia, and therefore lack professional training to engage in economic analysis. Additionally, no clear legal test has been set out in Shiva Shakti to balance between an individual’s strict rights under law and the nation’s economic interest. Increased litigation is expected due to the uncertainties introduced by Shiva Shakti, and this will hinder economic development. It is respectfully submitted that the judiciary’s promotion of economic growth and liberalisation in India might be better achieved by overhauling the operation of the court system instead.24

[1] [2017] 7 SCC 729

[2] Means-ends rationality is an assessment of the extent to which a course of action pursued by an agent is useful in helping it achieve its goals.

[3] Supra note 1.

[4] Interestingly, explanation 2 of Clause 6A defines a “new sugar factory” as one that is not operational, but has filed an IEM.

[5] Supra note 1 at para 22.

[6] Ibid at para 23.

[7] Ibid at para 28.

[8] Ibid at [34].

[9] Ibid at [37].

[10] Ibid.

[11] Ibid at [40].

[12] Ibid at [37].

[13] Ibid at [51].

[14] Margaret H. Lemos, “Interpretive Methodology and Delegations to Courts: Are “Common-Law Statutes” Different?”, in Shyam Balganesh ed, Intellectual Property and the Common Law (New York: Cambridge University Press, 2011) 89 at 90.

[15] The Sherman Antitrust Act, USC 15 §1(1890).

[16] Supra note 1 at [22].

[17] Ibid at [36].

[18] Ibid at [38].

[19] See R v Director of Public Prosecutions, ex parte Kebilene [1999] 2 AC 326 at 380; Lochner v New York [1905] 198 US 45 at 75; Galstaun v Attorney-General [1981] 1 MLJ 9 (HC).

[20] Lochner v New York [1905] 198 US 45 at 75.

[21] For instance, consider whether the Court would permit the privatization of a state-owned enterprise, knowing full well that the disabled at that company will be laid off, even though the current management of the state-owned enterprise have done a terrible job in increasing the company’s profits.

[22] Nirmala Ganapathy, “33 million cases waiting to be heard in Indian courts” The Straits Times (28 April 2016), online: <http://www.straitstimes.com/asia/south-asia/33-million-cases-waiting-to-be-heard-in-indian-courts>.

[23] Ibid.

[24] As Mr Lee Kuan Yew remarked in his speech delivered at the University of Singapore Law Society’s annual dinner in 1962, “The acid test of any legal system is not the greatness or grandeur of its ideal concepts, but whether in fact it is able to produce order and justice in the relationships between man and man and between man and the State”.

A PDF version of this article is available for download here.