1 This appeal arose out of certain demolition works performed by Esthetix Design Pte Ltd (“Esthetix”) on the respondents’ property. The respondents had hired Esthetix to demolish the existing dwelling house on their property and to construct another in its place. In the course of the demolition, debris fell on the appellants’ property, causing damage. The District Judge found that Esthetix was negligent and there was no appeal against that decision. However, he found that the respondents were not liable because (a) Esthetix was an independent contractor, (b) the respondents had exercised reasonable care in the selection of Esthetix, and (c) the demolition works in question were not “ultra-hazardous” and therefore did not give rise to a non-delegable duty of care. Dissatisfied, the appellants appealed against all three of the District Judge’s findings.

2 As this case touched on a few important points of law, policy and principle, one of which was the fact that this appeared to be the first time that the so-called “ultra-hazardous exception” had been considered in Singapore at length, Mr Keith Han was appointed as amicus curiae to assist this court. I record my appreciation for his submissions on the scope and applicability of the “ultra-hazardous exception”, which were succinct, thorough, and well-researched. I derived considerable assistance from them as well as the submissions put forward by Mr N Sreenivasan SC, and Ms Os Agarwal, both of whom argued their respective clients’ cases forcefully.

3 After careful consideration of the arguments presented, I was not persuaded that the District Judge had erred and I therefore dismissed the appeal. I concluded that the respondents were not vicariously liable as Esthetix was an independent contractor. I was also not persuaded that there was any basis for concluding that the respondents bore any personal liability for the damage that had been caused as they neither (a) failed to exercise due care in the selection of Esthetix as their contractor nor (b) did they owe the appellants a non-delegable duty arising out of the performance of the demolition works. I now set out the detailed grounds for my decision.

4 The facts were largely undisputed and were set out comprehensively in the District Judge’s grounds of decision, which were reported as Ng Huat Seng and another v Munib Mohammad Madni and others [2015] SGDC 315 (“the GD”). I will only set out the facts which are germane to this appeal. The parties are owners of neighbouring detached properties. The appellants’ house was the lower: the ground level of their house was two metres lower than the house which was on the respondents’ property, which lay further up the slope. Their properties were separated by a wall located on the boundary between the adjoining lands (“the boundary wall”). The building lines of each house were three metres away from the boundary wall, which meant that the distance between the building lines of the parties’ houses was six metres. ¹

5 The respondents had purchased their property in 2010 with the intention of demolishing the existing house and building another in its place (“the works”). The respondents hired Esthetix, a locally incorporated company that held a Class 2 General Builder’s Licence from the Building and Construction Authority (“BCA”), as their builder to carry out the works. The first respondent deposed that this appointment was made on a so-called “turnkey” basis, by which it was meant that Esthetix – as the main contractor – assumed carriage of the entire project and was responsible both for designing the house and for building it, engaging such subcontractors and applying for such approvals as might be required. He explained that this differed from what he called the “traditional approach”, where the owner would engage a team of professionals to design the house and obtain the necessary approvals before calling for a tender and appointing a main contractor. ²

6 Esthetix appointed a number of professional consultants on the project, who were: ³

(a) BDL Group Architects (“BDL”), which provided architectural services. Mr Wang Chun Jye of BDL was the development’s qualified person for architectural work.

(b) TH Chuah & Partners LLP (“THC”), which provided civil and structural engineering services. Er Lee Yen Fong from THC was the development’s qualified person for structural work.

(c) Tenwit Consultants Pte Ltd for the geotechnical engineering services.

7 Approval from the BCA was sought and obtained on 27 June 2011. ⁴ On 5 September 2011, while demolition works were taking place, some debris from the respondents’ property fell on the boundary wall, damaging it. Some of the debris also rebounded off the boundary wall and into the appellants’ property. Among other things, the falling debris broke several window panes, damaged several air-conditioning condensing units located at the exterior of the appellants’ house, and damaged the integrity of the boundary wall. The cost of repairing the damage caused was eventually assessed by the District Judge to be $136,796 (see the GD at [74(b)]).

8 On 22 May 2012, the appellants commenced District Court Suit No 1426 of 2012, naming the respondents and Esthetix as joint defendants. In their statement of claim, the appellants pleaded that the demolition works were “particularly hazardous and/or extra-hazardous” and that the respondents were personally liable for failing to “exercise reasonable care to avoid or prevent the damage and loss”. It was also pleaded that the appellants had failed to exercise reasonable care in the appointment of Esthetix. ⁵ In their defence, the respondents denied that the works had been carried out under their “control, supervision and/or management”. Instead, they pleaded that Esthetix was an independent contractor which they had taken reasonable care in selecting and entrusting the performance of the aforementioned works to. ⁶

9 The District Judge first considered whether Esthetix was a servant (or, in modern parlance, an employee), in which event the respondents would be held vicariously liable for its actions, or an independent contractor, in which event they would not be vicariously liable (see the GD at [21]–[22]). After reviewing the case-law, the District Judge held that there were two factors which pointed strongly towards the conclusion that Esthetix was an independent contractor.

10 First, he found that the respondents exercised little control over the manner in which Esthetix was to carry out its work. Among other things, the District Judge pointed to the fact that the appointment was made on a “turnkey” basis and to the fact that Esthetix enjoyed “significant autonomy when selecting and appointing the sub-contractors” with whom it contracted directly (see the GD at [27]–[29]). Second, he found that it was clear that Esthetix had taken on the project as part of its business for its own account. Apart from the fact that Esthetix had entered into contracts with the subcontractors in its “own name”, the District Judge also noted that Esthetix had charged the respondents goods and services tax. For these reasons, among others, the District Judge concluded that Esthetix was “an independent contractor, and not a servant” of the respondents (see the GD at [33]–[35]).

11 Second, the District Judge considered whether the respondents had been negligent in the selection of Esthetix as its contractor. After examining the facts and circumstances in their entirety, he held that the respondents had not fallen short of the standard of care expected of them in the selection of an independent contractor.

12 Chief among his reasons was the fact that Esthetix held a “Class 2” general builder’s licence from the BCA. This was significant, he held, because the grant of such a licence was contingent on satisfaction of the statutory requirements in the Building Control Act (Cap 29, 1999 Rev Ed), which required applicants to show that the execution of any building works would be supervised by a person with the relevant technical experience. He also noted that the respondents had solicited the opinion of their friends and sought the advice of their architect, BDL, before confirming Esthetix’s engagement (at [41]) and that there was no evidence that Esthetix had breached any regulations or was unsuitable to undertake the works for any reason (at [37] and [38]). In the circumstances, he found that it was “unobjectionable” for the respondents to have left the project in the hands of Esthetix and the qualified persons Esthetix appointed on a “turnkey” basis. As the respondents were laypersons, the District Judge held that it would be “highly unrealistic” to expect them to personally supervise the works carried out (at [40]).

13 Last, the District Judge considered if the works were “ultra-hazardous” and thus gave rise to a non-delegable duty of care. He noted that the “ultra-hazardous exception” had been the subject of extensive academic and judicial criticism (at [47]–[49]). Relying heavily on the decision of the English Court of Appeal in Biffa Waste Services Ltd and another v Maschinenfabrik Ernst Hese GmbH and others [2009] 3 WLR 324 (“Biffa Waste”), he held that the exception should be kept “as narrow as possible” and “should be applied only to activities that are exceptionally dangerous whatever precautions are taken” (see the GD at [51], citing Biffa Waste at [78]). Applying that approach to the case, he held that the works did not cross the threshold to being “ultra-hazardous”.

14 The District Judge noted that there was no proof that any explosives or any other inherently dangerous procedures were used (at [53]). Demolition works, while unquestionably dangerous, were not statutorily regarded as being so exceptionally dangerous that a separate permit had to be obtained in order for it to be carried out (at [55]). He rejected the appellants’ contention that the danger arose due to the proximity of the houses. These matters, he explained, should not be taken into account because they related to the surrounding circumstances, rather than to the question of whether the works were so dangerous per se that they gave rise to a non-delegable duty of care (at [59]). In the premises, he concluded that the respondents were entitled to rely on the “independent contractor defence” and dismissed the claims against them.

15 The parties agreed that the following three issues arose for decision: ⁷

(a) Whether Esthetix was an independent contractor;

(b) Whether the respondents had exercised due care in selecting and appointing Esthetix as their builder; and

(c) Whether the demolition works were “ultra-hazardous” and therefore gave rise to a non-delegable duty of care.

16 The first issue concerns the question whether the respondents are vicariously liable for Esthetix’s negligence. It is crucial to note that vicarious liability is a form of derivative liability. What this means is that an employer who is vicariously liable is free of the specific fault which occasioned the tort (otherwise he would be liable as a primary tortfeasor) but the law, for reasons of policy, nevertheless holds him legally responsible to make good the damage caused by another: see Skandinaviska Enskilda Banken AB (Publ), Singapore Branch v Asia Pacific Breweries (Singapore) Pte Ltd and another and another appeal [2011] 3 SLR 540 (“Skandinaviska”) at [76]–[80]). The independent contractor inquiry is relevant because it is settled law that an employer is not vicariously liable for the negligence of an independent contractor: see Management Corporation Strata Title Plan 2297 v Seasons Park Ltd [2005] 2 SLR(R) 613 (“Seasons Park”) at [37].

17 The second issue relates to the respondents’ alleged liability for failing to exercise reasonable care in the appointment of an independent contractor. I shall refer to this as “negligent selection liability”. Negligent selection liability is personal and it is primary in the sense that the legal responsibility of the employer flows from his own actions in failing to exercise reasonable care in the selection process and not from the negligent acts of another. If an employer does not exercise proper care in the appointment of the independent contractor then he would be liable for his own lack of care even though he might not be vicariously liable for torts committed by the independent contractor (see Seasons Park at [37]).

18 The third issue relates to whether the respondents were in breach of a non-delegable duty of care. The common law has long recognised that there exist certain types of cases where, either because of the nature of the act in question or because of the character of the relationship between the defendant and the victim, a defendant may be liable for the torts performed by another, even if that other might be an independent contractor: see Woodland v Swimming Teachers Association and others [2014] AC 537 (“Woodland”) at [5]–[7]. These are commonly referred to as “non-delegable duties”. One recognised instance where a non-delegable duty arises is where the act in question is “extra-hazardous” (see Seasons Park at [38]). Whether the works fall within this description is what the third issue is principally concerned with.

19 As Chan Seng Onn J recently emphasised in Management Corporation Strata Title Plan No 3322 v Mer Vue Developments Pte Ltd and others (King Wan Construction Pte Ltd and others, third parties) (“Mer Vue”) [2016] 2 SLR 793, it is incorrect to say that non-delegable duties are an “exception” to the rule that vicarious liability cannot arise out of the tort of an independent contractor. Rather, non-delegable duties are “primary and personal” [emphasis in original], and the breach of a non-delegable duty generates a distinct basis for imputing personal liability on the employer (see Mer Vue at [16]). It was stressed that liability for the breach of a non-delegable duty is not to be taken as a “disguised form of vicarious liability where secondary liability is still imposed on the employer for its independent contractor’s tortious acts” [emphasis added] (likewise at [16]).

20 Taken together, these three issues map out the terrain of liability which the respondents were alleged to be confronted with. They relate to distinct but inter-related grounds for imputing liability on the respondents. I propose to take the issues in that order, for that was how it was argued before me. I will examine the detailed arguments presented by the parties as I go along but a brief synopsis will suffice for now.

21 Mr Sreenivasan submitted that Esthetix was not an independent contractor so the respondents are vicariously liable (qua employers) for the negligence of Esthetix. In the alternative, he contends that the respondents are personally liable because they had (a) failed to exercise reasonable care in the selection of Esthetix as their builder and (b) because the works were “ultra-hazardous” and therefore gave rise to a non-delegable personal duty of care. ⁸ Ms Agarwal, on the other hand, argued to the contrary on all three issues. She contended that Esthetix was an independent contractor which the respondents had exercised reasonable care in appointing and that the works Esthetix had been appointed to perform were not ultra-hazardous per se. She therefore submitted that there was no basis for holding the respondents liable, whether vicariously or personally. ⁹

22 Mr Sreenivasan did not appear to dispute the District Judge’s application of the law as it stood. Instead, his primary argument was that the time had come for this court to adopt a different test for determining when the “defence of independent contractor” should be available to an employer. Indeed, he went further than that, for he effectively argued that an entirely new approach should be taken towards questions of vicarious liability altogether. He contended that the court should move beyond what he termed the “traditional test of employment” as a basis for the imposition of vicarious liability (which, concomitantly, suggests that vicarious liability should not be imposed for the acts of an independent contractor) towards a more policy-centric approach. What mattered most was whether the policy considerations operating in this area of the law – (a) effective compensation for victims; (b) loss-spreading; and (c) deterrence of future harm – behoved the imposition of vicarious liability. He argued that these three policy considerations justified the imposition of vicarious liability in this case. ¹⁰ Alternatively, he argued that it would be “fair and just” to hold the respondents vicariously liable because there was a “close connection” between the risks posed by the works Esthetix were hired to perform and the harm suffered by the appellants. ¹¹ In support of his submissions, Mr Sreenivasan relied heavily on the decision of the UK Supreme Court in Various Claimants v Catholic Child Welfare Society and others [2012] 3 WLR 1319 (the “Christian Brothers case”).

23 With respect, I could not agree with his submissions. To explain why, I find it necessary to discuss the law of vicarious liability more generally before coming back to the issue of whether Esthetix is an independent contractor. I shall first examine the judgment of the UK Supreme Court in the Christian Brothers case in some detail because of the importance it played in the parties’ submissions. I then consider how the independent contractor argument may be understood in light of my analysis of the Christian Brothers case before examining the arguments presented by Mr Sreenivasan in detail.

24 The question in the Christian Brothers case was whether a Roman Catholic teaching order (“the Institute”), an international unincorporated association, was vicariously liable for the sexual abuse committed by members of its order who had taught at a boys’ residential school in England. The school was managed by another organisation which ran the school and concluded contracts of employment with the teachers who taught there. The English High Court and Court of Appeal both concluded that this organisation was vicariously liable but that the Institute was not. There was only an appeal against the finding that the Institute was not vicariously liable. On appeal, the UK Supreme Court reversed the decisions of the lower courts and held that the Institute was vicariously liable.

25 Lord Phillips of Worth Matravers, with whom the rest of the court agreed, explained that any discussion of vicarious liability classically invited a two-stage inquiry (at [19]): (a) first, whether there a true employer/employee relationship between the defendant and the tortfeasor; (b) second, whether the tortfeasor was acting “in the course of his employment” when he committed the tortious act. Collectively, these set out the two cumulative conditions that had to be satisfied in order for a defendant to be held vicariously liable for the conduct of another. For ease of exposition, I have, drawing on the judgment of Lord Reed JSC in Cox v Ministry of Justice [2016] 2 WLR 806 (“Cox”) at [2], reformulated these two stages in more general terms (stripped of the language of employment – the significance of which will soon be clear):

(a) First, the relationship between the tortfeasor and the defendant must be of a type which is capable of giving rise to a finding of vicarious liability.

(b) Second, the conduct of the tortfeasor must possess a sufficient connection with the relationship between the tortfeasor and the defendant such that it may be said that vicariously liability may arise.

26 In relation to the first issue, Lord Phillips held that while the doctrine of vicarious liability had developed in the context of an employer-employee relationship, it was no longer so limited. Thus, the fact that the perpetrators were not employees of the Institute was not an insuperable barrier to a finding of vicarious liability. At [35], he put the point in this way:

These five factors, he later explained at [47], set out those “incidents of the relationship between employer and employee that make it fair, just and reasonable to impose vicarious liability”. Where they are present, it may properly be said that the relationship is one which is “akin to that between an employer and an employee” and is therefore one which may give rise to vicarious liability (likewise at [47]).

27 On the facts, he held that the relationship between the Institute and its members satisfied this test. Among other things, he pointed to the fact that the Institute conducted its affairs as a corporate body and that it directed the emplacement of its members in the schools within its worldwide network. All of the perpetrators had been sent to the school at the direction of the Institute and one of them served as its headmaster. Furthermore, while the members of the Institute were bound to it not by contract, they had taken religious vows in which they swore to pay their earnings to the Institute and to follow its precepts, both in the conduct of their lives and in the manner in which they conducted themselves as teachers. In the circumstances, he concluded that the relationship between the Institute and its members was, if anything, closer than that of one between an employer and its employees and was clearly capable of giving rise to vicarious liability.

28 In relation to the second issue, Lord Phillips explained that the traditional approach was that an employer would be vicariously liable for all acts of an employee that were done “in the course of the employment” (at [62]). This clearly embraced situations involving the negligent performance of an authorised act, but the difficulty lay in cases where an employee had intentionally committed a tort. In such a situation, the modern position in the English authorities was that vicarious liability could be found where there was a “close connection” between the tort and the relationship between the defendant and the tortfeasor. Elaborating, he explained that this grew out of the modern approach towards vicarious liability, which saw matters through the prism of the theory of “enterprise risk”: since employers, through engaging in business, had put the risk of tortious acts arising out into the world, it was fair and just that they should be liable when such risk eventuated (at [75]). At [86], he put matters in the following terms:

29 On the facts, Lord Phillips held that this requirement had also been satisfied. The central mission of the Institute was the provision of Christian education to boys under its care, and it was in the furtherance of this mission that the perpetrators were sent to teach at the school. Further, each of the perpetrators had only come to teach at the school because of their status as members of the Institute, and once they were there, the victims – who came from difficult backgrounds and were particularly vulnerable as a result – were placed in their charge. In conclusion, the perpetrators’ abuse of these children, while a complete abnegation of their duties to the Institute and a gross abuse of their positions, were acts which bore a close connection with their relationship with the Institute and were therefore acts in respect of which the Institute could be held vicariously liable.

30 The Christian Brothers case is instructive both because it provides a comprehensive re-statement of the current position of English law on the subject of vicarious liability and because of the two-stage approach which it introduced. It was recently affirmed in a pair of decisions handed down by the UK Supreme Court early this year (see Cox at [2], Mohamud v Wm Morrison Supermarkets plc [2016] 2 WLR 821 at [1]). However, it must be remembered that the facts of the Christian Brothers case were quite unique for two reasons. First, the perpetrators of the abuse, while members of the Institute, were not its employees. Second, the sexual abuse was clearly something that the perpetrators were specifically enjoined from doing. It was against that background that the two-stage approach was propounded as a basis for identifying (a) the circumstances in which vicarious liability may be imposed outside relationships of employment and (b) to reflect the modern consensus about the responsibilities of businesses for the risks created by their activities, even if the harms that flow from these risks do not, strictly speaking, arise as a result of acts or omissions taken by the employees of these businesses in the course of their employment (see Cox at [29]).

31 At a later part of the same paragraph of Cox, Lord Reed JSC explained the significance of the Christian Brothers case and the two-stage approach which it introduced in the following terms (likewise at [29]):

32 I have emphasised the phrase “maintain previous levels of protection” to stress that the two-stage approach was never intended to inaugurate a radical change in the law of vicarious liability, but to systematise and update it in the light of modern business realities. The two-stage approach is valuable because it identifies the two core requirements that must be satisfied in all cases of vicarious liability; as an analytical tool, it really comes into its own in marginal cases, where it provides a principled and clear framework for use in determining if vicarious liability should arise outside relationships of employment, which had traditionally been the preserve of vicarious liability. However, as Lord Reed JSC emphasised, the two-stage approach should not be taken to have rendered the old case-law otiose or to have sanctioned a significant expansion in the ambit of vicarious liability. This is a significant point which I shall return to in the next section.

33 As I noted at [16] above, it is well-established that vicarious liability does not attach to acts performed by independent contractors. The reason for this, to put it in terms of the two-stage approach, is that if the tortfeasor is an independent contractor the requisite employer-employee relationship is absent and there can be no basis for holding the employer vicariously liable. It is only in this sense that a plea that the tortfeasor is an independent contractor may be said to be, as Mr Sreenivasan put it, a “defence” to vicarious liability: see BNM (administratrix of the estate of B, deceased) on her own behalf and on behalf of others v National University of Singapore and others and another appeal [2014] 4 SLR 931 (“BNM”) at [16]. ¹²

34 The reason for this blanket exclusion of independent contractors is because there is generally no justification, either in policy or in principle, for the imposition of vicarious liability for the acts of independent contractors. In Skandinaviska, Chan Sek Keong CJ, delivering the judgment of the court, explained at [76] that two important policy considerations underpinned the doctrine of vicarious liability, namely: (a) effective compensation for the victim and (b) deterrence of future harm by encouraging employers to take steps to reduce the incidence of accidents or tortious behaviour by their employees through efficient organisation and supervision. However, neither of these aims would be achieved through the imposition of vicarious liability for the acts of independent contractors.

35 Independent contractors are hired in a variety of situations and engaged to perform a variety of tasks and it is very often the case that the hirers are individuals, rather than persons who are carrying on a business. Often, these hirers do not possess the financial wherewithal to make good the losses sustained by the torts of these contractors. Further, they are often in no position to exert any meaningful control over the activities of independent contractors since they are not part of their organisation. To make the hirers vicariously liable for the acts of independent contractors would neither serve the aim of victim compensation nor that of effective deterrence.

36 But even if it could be shown that the aims of victim compensation and deterrence could be achieved through the imposition of vicarious liability on these hirers, it would not be enough. As Chan CJ cautioned at [81] of Skandinaviska, “[t]he courts are… neither welfare agencies nor workplace safety regulators”. The fact that a person can easily bear a loss does not, without more, justify the conclusion that the law ought to make him do so. Tort law is still primarily a system of norms of personal responsibility and a principled moral basis for the imposition of liability is required (see Peter Cane, “The Changing Fortunes of Rylands v Fletcher” (1994) 24 WAL Rev 237 at 243). This moral basis can be found in the concept of “enterprise risk”, which Chan CJ alluded to at [77] of Skandinaviska, where he wrote, “a person who employs another to advance his own interests and thereby creates a risk of his employee committing a tort should bear responsibility for any adverse consequences resulting therefrom.”

37 The concept of an “enterprise risk” rose to prominence after it was discussed in the decision of the Canadian Supreme Court in Bazley v Curry [1999] 2 SCR 534 at [22] and it now dominates modern discussions on the subject of vicarious liability (see Cox at [23]–[24] per Lord Reed; the Christian Brothers case at [86] per Lord Phillips). It is not just a unifying idea or grand theme but the normative foundation for the doctrine of vicarious liability. As Rix LJ put it in Viasystems (Tyneside) Ltd v Thermal Transfer (Northern) Ltd and others [2006] 2 WLR 428 at [55], “those who set in motion and profit from the activities of their employees should compensate those who are injured by such activities even when performed negligently.” In short, where the incidence of benefit falls, so, too, should the risk. Since employers, through engaging in business, have put the risk of tortious acts arising out into the world, they should be liable when the risk eventuates.

38 The point, for present purposes, is that the imposition of vicarious liability for the acts of independent contractors can never be justified on the basis of the theory of enterprise risk. By definition, an independent contractor is one who is not part of the hirer’s organisation and is instead a person who is “performing services as a person of business on his own account” (see the GD at [24]). The independent contractor carries on business for his own benefit, and thus, any risks of harm arising from the independent contractor’s conduct (the enterprise risk) should properly fall on the independent contractor alone. As I noted above, even the UK, which has embraced the policy-driven approach, has not countenanced the imposition of vicarious liability for acts performed by independent contractors (see [32] above).

39 For these reasons, I agreed with Mr Sreenivasan that the modern approach towards the question of vicarious liability has been to move away from a rigid adherence to labels (“employer-employee”; “course of employment” etc.) towards a more open-textured analysis grounded in a consideration of the policy objectives underpinning the doctrine of vicarious liability (see Skandinaviska at [82]–[83]). However, I did not agree that this meant that the distinction between employees and independent contractors should be jettisoned entirely. As I explained above, considerations of principle and policy continue to militate against the imposition of vicarious liability and the same outcome (ie, that the respondents should not be held vicariously liable for Esthetix’s negligence if it could be shown that Esthetix was an independent contractor) would also be reached on an application of the two-stage approach set out in the Christian Brothers case.

40 I was also not persuaded by Mr Sreenivasan’s alternative argument, which was that the “close connection” test could be used to justify the extension of vicarious liability. In my judgment, this argument puts the cart before the horse. In order for vicarious liability to arise, two cumulative conditions must be satisfied: (a) the relationship must be capable of giving rise to vicarious liability and (b) the tortious act must bear a sufficient connection with the said relationship. Each limb of the test serves a different purpose: The former defines the circumstances in which vicarious liability may arise; the latter operates to limit the ambit of vicarious liability within the confines of that relationship. Here, the appellants’ argument fails at the first hurdle – the relationship between the respondents and Esthetix is not of a type which can give rise to vicarious liability.

41 As I noted at the outset (see [22] above), Mr Sreenivasan did not dispute the District Judge’s identification of the applicable legal test for determining if a person is a servant or independent contractor or that, upon the application of this existing test, Esthetix was an independent contractor. Because of this, all that really need be said is that I agreed with the District Judge that (a) the applicable test is that which was that set out by the Court of Appeal at [29] of BNM: ie, “whether the contractor was performing services as a business on his own account”; and (b) that on the application of this test, Esthetix was clearly an independent contractor. In my assessment, the most important factors are that Esthetix: ¹³

(a) concluded contracts with the various consultants and sub-contractors in its own name (see the GD at [29]);

(b) hired its own employees and was solely responsible for their management and supervision; ¹⁴

(c) took out insurance in its own name (see the GD at [71]); ¹⁵

(d)  maintained a separate account from the respondents and regularly received lump sum payments from the respondents which it retained as profits (see the GD at [33]).

42 For these reasons, I upheld the District Judge’s decision that the respondents are not vicariously liable for Esthetix’s negligence because the latter was an independent contractor.

43 Generally speaking, all who seek to carry out tasks through the agency of an independent contractor (I use the expression “agency” in its non-legal and non-technical sense) owe a duty of care to those who might be affected by the actions of the independent contractor to exercise reasonable care and skill in the selection of a suitable contractor. In this case, the parties did not dispute that such a duty existed. Instead, they only disagreed on whether it had been breached on the facts. ¹⁶

44 Mr Sreenivasan accepted that the respondents took some steps to satisfy themselves that Esthetix would be a suitable contractor – these steps were set out in the affidavit that the first respondent filed in this suit, and they were succinctly summarised in the District Judge’s GD (see [12] above). However, Mr Sreenivasan submitted that the steps taken were insufficient and that the respondents ought to have done more. Among other things, he said that the respondents ought to have examined Esthetix’s Accounting and Corporate Regulatory Authority (“ACRA”) records, and taken steps to find out how much experience Esthetix had with the conduct of demolition works. In particular, he stressed that it was insufficient for the respondents to have relied on the testimony of their friends, who were laypersons like themselves, and that they should have sought the views of persons who were familiar with the industry before appointing Esthetix as their builder. ¹⁷

45 The principal difficulty with this submission is that it rests on the premise that had the respondents done as Mr Sreenivasan says they ought, they would have discovered that Esthetix was unsuitable and not have selected it as their builder. The issue is one of causation. In order for the appellants to make out their case against the respondents on the ground of negligent selection liability, it would not be enough to show that the respondents had been derelict in their selection of a contractor; they must also show that the respondent’s dereliction of their duty of careful selection had caused them loss. In other words, the appellants must show that the respondents would not have selected Esthetix if they had exercised reasonable care in the selection process. For the reasons which follow, the appellants had not, in my judgment, proved this.

46 At the time it was selected by the respondents as their contractor, Esthetix (a) had the requisite Class 2 Builder’s licence; (b) had a clean safety record; (c) was recommended by BDL – a firm of professional architects – and by the respondents’ friends, both of whom had worked with Esthetix before. In short, as Ms Agarwal pointed out, there was no evidence to suggest that Esthetix was an unsuitable for appointment. ¹⁸ The only red flag which Mr Sreenivasan was able to point to was the fact that the performance of demolition works was not listed as one of the principal activities performed by Esthetix on its ACRA profile. ¹⁹ However, as the District Judge rightly pointed out, this was merely information which was provided for the purposes of Esthetix’s incorporation (which took place in 2007), and was not particularly probative as to whether Esthetix had the requisite expertise in the area of demolition works (see the GD at [39]). ²⁰

47 The same went for Mr Sreenivasan’s point that the respondents ought to have consulted with persons who were familiar with the industry. As a preliminary point, I note that the respondents did in fact seek the views of persons in the industry, for they sought the views of BDL, which recommended Esthetix. But even putting that aside, the pertinent point is there was no evidence that further industry inquiries would have made any difference. The appellants did not lead any evidence to the effect that Esthetix had a bad safety record, that it had acquired a reputation for poor performance in the industry, or that there was otherwise anything irregular which would have been turned up in an inquiry that would have made a difference to the appointment. Shortly put, there was simply no evidence upon which the court might conclude that the inquiries would have resulted in answers that would have been unfavourable and led to an alternative appointment.

48 A parallel may be drawn with Hong Cassley and others v GMP Securities Europe LLP and another [2015] EWHC 722 (QB) (“Cassley”). The plaintiffs in that case were the dependents of a man who had died in an air-crash while on company business. It was ascertained that the crash had occurred due to pilot error. The plaintiffs sued the company, arguing that they had been negligent in not making more inquiries. It was pleaded that had such inquiries been made, the company would not have allowed their employee to board the plane and would instead have put him on a flight managed by a more competent carrier, averting the disaster. This argument was rejected by Coulson J. He found that even if the company had made the inquiries which the plaintiffs said they ought, this would not have made a difference. The company would still have been told that the carrier selected had been recommended by other pilots, that it had the requisite certification, that it had successfully plied that route before, and that the second defendant had already used the carrier’s services successfully before (at [250]). In essence, the company would not have heard anything to doubt the reliability of the carrier and led them to forbid their employee from boarding the plane. In the circumstances, the claim against the company faced a “causation gap” which was “unbridgeable” and had to fail (at [236]).

49 Likewise, it seemed to me that even if the respondents had done whatever Mr Sreenivasan submitted they ought to have done (and assuming further that their failure to do so constituted a breach of duty), there was no evidence that the respondents would have chosen any differently or that they would have had any reason to have chosen differently. Indeed, there was no evidence that the respondents were considering alternative contractors besides Esthetix. Without proof that the respondents would have made an alternative selection, the appellants could not succeed in establishing that the respondents’ alleged breach of duty led to their loss and their claim failed.

50 The case of Michael John Bottomley v Todmorden Cricket Club [2003] EWCA Civ 1575 (“Bottomley”), which Mr Sreenivasan cited in support of his argument, is readily distinguishable. In Bottomley, the defendant cricket club hired a stunt-team to perform a pyrotechnic display on their premises. The stunt-team engaged the assistance of the plaintiff, a volunteer, to help them prime the mortars by filling them with gunpowder. As the plaintiff was priming the second mortar, its contents exploded and he was severely injured. The plaintiff then brought suit against the cricket club, alleging that they had failed in their duty to take reasonable care in the selection of a suitable contractor. At first instance, the judge found as a fact that the stunt team were “inexperienced and largely ignorant of basic safety requirements for the discharge of pyrotechnics” and, further, that this fact could be discovered if reasonable care were taken to inquire into its credentials (at [25]). This finding was upheld by the English Court of Appeal, which commented that the performance of such basic checks would, in all likelihood, have led to either the cancellation of the event or an obligation on the stunt team to show the cricket club its safety plan, which would have required it to pay proper attention to the plaintiff’s safety. These steps would probably have prevented the accident (at [40] and [48]).

51 The difference between Bottomley and the present case is plain. Bottomley concerned an independent contractor in respect of which proper checks would have revealed to be (a) manifestly unsuitable for the task assigned and (b) led to the cancellation of the event or at least led the cricket club to insist that a safety plan be furnished. For that reason, the cricket club’s failure to conduct the requisite checks was causally linked to the plaintiff’s injury. The causal link was absent in this case. Even if the respondents had conducted the checks that the appellants said they should, they would not have uncovered anything to suggest that Esthetix was unsuitable for appointment or given it reason not to have selected Esthetix. Without causation, there cannot be a complete tort of negligence. In my judgment, this is fatal to the appellants’ case and I therefore dismissed the appeal on the second issue.

52 In the circumstances, there is no need for me to examine in detail whether the respondents had breached their duty of care. All I will say is that I was not persuaded that they had, for the reasons given by the District Judge. Whether there has been a breach of duty depends crucially on the standard of care to be applied. The standard to be applied is that of a reasonable person in the circumstances of the defendant (see Blyth v The Company of Proprietors of the Birmingham Waterworks (1856) 11 Ex 781 per Alderson B). The salient facts of the present case were these. The respondents were laypersons who desired to construct their own home. To that end, they employed the “turnkey” approach, which affords a contractor a great deal of latitude in the construction process. While it was put to the respondents that they only had chosen this method because it was cheaper (they denied this), it was not challenged that this was an accepted industry practice at the time and a “common choice for homeowners in Singapore”. ²¹

53 The respondents then took the following steps before appointing Esthetix as their builder:

(a) They consulted their friends who had worked with Esthetix and had received a favourable report. ²²

(b) They ascertained that Esthetix had the requisite Class 2 licence from the BCA. ²³

(c) They obtained an assurance that Esthetix would obtain the necessary approvals and take the requisite safety precautions before proceeding. ²⁴

(d) They relied on their personal interactions with Esthetix, whom they had consulted when they viewed various properties with a view towards purchasing one. ²⁵

54 When I considered the matters in their totality, I was satisfied that the respondents had done that which could be expected of reasonable persons in those circumstances and that they were not negligent in choosing Esthetix as their builder.

55 I now turn to the final issue, which concerns the question whether the respondents bore a non-delegable duty of care to ensure that Esthetix took reasonable care in the performance of the works. This is a difficult area of the law, and one in which much has been written of late. I propose to first say something about the general nature of non-delegable duties before examining the detailed arguments put forward by the parties.

56 The expression “non-delegable” is apt to mislead insofar as it suggests that the delegation of a task to another is usually accompanied also by a delegation of the duty of care associated with the performance of that task, and that the cases in which non-delegable duties arise are an exception to this. This is not correct. Strictly speaking, a task can be delegated, but a duty of care cannot. Take, for example, the case of M’Alister (or Donoghue) (Pauper) v Stevenson [1932] AC 562 (“Donoghue v Stevenson”). The manufacturer of the ginger beer owed a duty of care to his consumers to ensure that he took reasonable care to manufacture products which were safe for consumption, including by taking steps to supervise the manufacturing process. To that end, he could have hired third party safety or quality control inspectors to watch over the production process to ensure that no foreign substances made its way into the product. If he did so, the manufacturer might have delegated the task of supervision to the inspectors, but the duty of ensuring that reasonable care was taken would still reside with the manufacturer: see Glanville Williams, “Liability for Independent Contractors” [1956] CLJ 180 at 184.

57 So what is the law of non-delegable duties concerned with? To answer this, one has to first remember that the general rule is that liability in tort is generally contingent on personal fault. A defendant is not ordinarily personally liable for the acts of others, as he – not being the one who performed the act – usually holds no relevant duty to ensure that the act was performed safely. Thus, a person who hires an independent contractor to do X ordinarily holds no duty of care in respect of the safe performance of X per se; instead, his duty of care is to select a reasonably competent contractor, for the only act he has performed is the hiring of the contractor, and it is only in respect of this act of hiring that he holds a duty of care (this relates to the notion of “negligent selection liability” discussed above). The contractor, who actually does X, is the one who bears a duty of care to do X with reasonable care. If the contractor performs X negligently, there is usually no basis for holding the hirer responsible.

58 One exception to the above is where the hirer is vicariously liable. In this case, secondary liability for the tort is imputed to the defendant by law. This is a form of derivative, not personal, liability and it does not apply where the person to whom the task is delegated is an independent contractor (see [33] above). Another exception, which is what this section is concerned with, is where the defendant holds a non-delegable duty. A non-delegable duty of care is one in which “the duty extends beyond being careful [in the performance of one’s own acts], to procuring the careful performance of work delegated to others (see Woodland at [5] per Lord Sumption JSC). A breach of a non-delegable duty of care gives rise to personal, not derivative, liability. A duty of this sort is only “non-delegable” in the sense that it is no answer to a breach of a non-delegable duty to say that the act in question had been performed by another, for the careful monitoring of work performed by others is precisely the scope of a non-delegable duty.

59 In short, to describe a duty as “non-delegable” is to state a proposition about the nature and content of the duty of care (see the decision of the High Court of Australia in Leichhardt Municipal Council v Montgomery (2007) 233 ALR 200 (“Leichhardt”) at [6] per Gleeson CJ). Specifically, it is to say that the duty of care consists of a positive duty to ensure that a third party does something with reasonable care. Positive duties are unusual, but not unheard of. Positive duties may be imposed by statute (see, eg, Mer Vue at [27]), but they may also be imposed under the common law. The common law imposes a positive duty on persons in respect of the acts of third parties in some instances, usually where there is a special relationship between the defendant and the third party or where the defendant is responsible for creating a source of danger which may be triggered by the third party (see Smith v Littlewoods Organisation Ltd [1987] AC 241 (“Smith”) at 271–274 per Lord Goff).

60 A duty of this sort often amounts, in practical terms, to strict liability. Strictly speaking, the content of a non-delegable duty is to “see that reasonable skill and care” is exercised by an independent contractor in the performance of his task; it does not amount to a legal guarantee that no harm at all would result (see George Martin Hughes v John Percival (1883) 8 App Cas 443 at 446 per Lord Blackburn). However, since independent contractors are persons over whom their hirers exercise little control, a duty to “see that” an independent contractor takes reasonable care in the performance of his duties is “a duty to do the impossible” (see Leichhardt at [23] per Gleeson CJ). What is required of the hirer is a degree of diligence so unattainable that it leads, for all intents and purposes, to strict liability (see Burnie Port Authority v General Jones Pty Ltd (1994) 120 ALR 42 (“Burnie”) at 65 per Mason CJ).

61 Another feature of non-delegable duties is that there is no unified theory to explain their existence. In Woodland at [6], Lord Sumption JSC explained that there were two broad situations in which non-delegable duties may arise (cited with approval in Mer Vue at [21]). The first was what he described as a “large, varied, and anomalous class of cases” in which an independent contractor had been hired to perform a function which was “inherently dangerous or likely to become so in the course of his work” (at [6]). The second was a group of cases in which there existed an assumption of responsibility by virtue of the special character of the relationship between the defendant and the claimant (at [11]–[12]). Both Woodland and Mer Vue were concerned with the second category; here, we are concerned with the first.

62 The District Judge held that the respondents did not hold a non-delegable duty in this case because the works in question were not ultra-hazardous. However, Mr Sreenivasan challenged the District Judge’s finding that the respondents did not hold a non-delegable duty on two main grounds:

(a) First, he argued that the District Judge had misread the decision of the English Court of Appeal in Biffa Waste and had therefore erroneously concluded that the surrounding circumstances (in this case, the proximity of the two houses) could not be taken into account in determining whether an activity is ultra-hazardous. He submitted that had the surrounding circumstances been taken into account, the proper conclusion to be drawn was that the works were ultra-hazardous and thus gave rise to a non-delegable duty of care. ²⁶

(b) Second, and even if the present case could not be described as “ultra-hazardous” as it was traditionally understood, he contended that the respondents nevertheless owed a duty of care to see to it that Esthetix performed the demolition works safely, pursuant to the general principles of the law of negligence. ²⁷

63 Mr Sreenivasan’s two submissions may be understood in the following way. The first submission engages the traditional approach towards determining the existence of a non-delegable duty of care and it asks the question whether the present situation falls within the first of two scenarios described in Woodland. The second submission relies on the more general approach in the law of negligence and it asks the question whether, on an application of the framework set out in Spandeck Engineering (S) Pte Ltd v Defence Science & Technology Agency [2007] 4 SLR(R) 100 (“Spandeck”), a duty of care with the features I identified at [59]–[60] arises on these facts. I will take each sub-issue in turn.

64 I begin with that area of the law that governs the conduct of what has variously been referred to as “ultra-hazardous”, “extra-hazardous”, or “unusually dangerous” acts. As observed in Biffa Waste at [62], nothing turns on these differences in nomenclature and I shall, consistently with the parties’ usage, adopt the expression “ultra-hazardous”. In essence, the proposition is that there are certain acts which are so dangerous that the law says that any person who procures their performance cannot escape legal liability if harm is caused by their negligent performance, even if the act itself was performed by an independent contractor. I shall refer to this principle as the “doctrine of ultra-hazardous acts” instead of the “ultra-hazardous exception”, which was how the parties and the District Judge had referred to. For the reasons set out at [19] above, I do not think that the expression “exception” is apposite.

65 The locus classicus is the decision of the English Court of Appeal in Honeywill and Stein Limited v Larkin Brothers (London’s Commercial Photographers) Limited [1934] 1 KB 191 (“Honeywill”). The plaintiffs had hired the defendants to take photographs of the interior of a cinema. The defendants’ employee, a photographer, used a chemical flashlight to illuminate the interior of the cinema. This process involved the ignition of magnesium powder in a tray held above the lens. Upon ignition, the powder produced an intense heat and an incandescent flash of light. It was accepted that this act of ignition was dangerous if carried out near fabrics. The photographer negligently set up his camera too close to a curtain which was set on fire when the magnesium powder was ignited. Considerable damage was caused to the cinema in the ensuing conflagration and the plaintiffs, acting on advice, paid the owners of the cinema before suing the defendants for an indemnity. The defendants argued that no damages were recoverable because the plaintiffs had no legal liability to pay up as the defendants were independent contractors. At first instance the claim was disallowed but this was reversed on appeal.

66 Slesser LJ, who delivered the judgment of the English Court of Appeal, noted that the general rule was that an employer was not liable for the acts of an independent contractor. However, there was a competing principle, which was that there were cases where “a person causing something to be done, the doing of which casts on him a duty, cannot escape from the responsibility attaching on him of seeing that duty performed by delegating it to a contractor” (at 197). Even though he did not use that precise expression, it is clear that this is the language of non-delegable duties. One such case was where one was concerned with ultra-hazardous acts. At 200, he put the matter in the following terms:

67 This principle soon became an established part of English law, albeit one honoured more in the breach rather than the observance. To the best of my knowledge, this principle has never been applied in Singapore though its existence has been recognised, albeit obiter, in multiple judgments of our Court of Appeal: see, eg, Mohd Sainudin bin Ahmad v Consolidated Hotels Ltd and another [1990] 2 SLR(R) 787 at [10]; The “Sunrise Crane” [2004] 4 SLR(R) 715 at [30]; Seasons Park at [38]. In the circumstances, it is necessary for me to first consider the precise contours of the doctrine before applying it to the facts of this case.

68 The starting point for any modern discussion on this topic must be the decision of the English Court of Appeal in Biffa Waste, which both parties referred me to. ²⁸ The case concerned a fire which was caused at a recycling plant. The first defendant, MEH, was hired to design and build the plant for the plaintiffs. MEH subcontracted the works to HU which, in turn, sub-subcontracted the supply and installation of the “ball mill” (an integral part of the plant) to the second defendant, OT. OT itself hired P to perform certain welding works in a section of the ball mill. It was acknowledged that there was an element of danger involved in this as the ball mill was used for the processing of organic material and one part of it was filled with combustible material. For that reason, a number of precautions were instituted and these included: (a) the requirement that a continuous watch be kept after the welding had been completed in order that any smouldering fires might be detected and put out and (b) that the welding area was properly wetted down after welding. A fire was started in the course of welding and it caused extensive damage. It was acknowledged that HU and P were negligent and did not follow the necessary precautions but as both of them were insolvent, the plaintiffs sought to sue MEH and OT directly.

69 The English High Court rejected the claims against both MEH and OT and the plaintiffs only appealed against his decision in respect of OT’s liability. On appeal, the plaintiffs’ case was based on two grounds: (a) P’s welders were OT’s employees pro hac vice and therefore OT was vicariously liable for its negligence and (b) the works carried out by P – welding in the vicinity of combustible material – were ultra-hazardous and therefore gave rise to a non-delegable duty. The English Court of Appeal rejected both arguments. For present purposes, I will only concern myself with the latter.

70 Stanley Burton LJ, delivering the judgment of the court, noted that the decision in Honeywill had been the focus of extensive criticism, most of which had focused on the argument that the very concept of an ultra-hazardous activity was nebulous. Citing P.S. Atiyah, Vicarious Liability in the Law of Torts (Butterworths, 1967) (“Vicarious liability”), he noted that the doctrine had “produced some quite preposterous distinctions arising out of the difficulty of saying what is an inherently dangerous operation” (at [75], citing Vicarious Liability at 371). He observed at [73] that the House of Lords had, in Read v J Lyons & Company Limited [1947] AC 156 (“Read”), rejected as “impracticable” the notion that any coherent distinction could be drawn between acts which were merely dangerous and those which were so dangerous that special rules of liability should attach. He drew particular attention to the following passage from the speech of Lord MacMillan from Read at 172:

(Although the specific question before the House of Lords in Read was different – the argument in Read was whether strict liability attached to the performance of extra-hazardous acts and the court concluded that it did not – Stanley Burton LJ agreed with Professor Atiyah that at the core of it, the decisions in Honeywill and Read were irreconcilable (at [75]).)

71 Further compounding the problem, he held, was the practical difficulty involved in determining whether an activity was ultra-hazardous. He drew particular attention to the issue of precautions. He pointed out that many everyday activities could be classified as inherently dangerous unless proper precautions were taken. He gave the example of driving. It was ultra-hazardous to drive without keeping a proper lookout but not ultra-hazardous if proper precautions were taken. However, the precaution of keeping a lookout was such an intrinsic part of driving that it would be “irrational” to exclude it from consideration when considering the dangerousness of the activity (at [75] and [76]). However, it appeared that this (the exclusion of precautions from a determination of the dangerousness of an act) was precisely what Slesser LJ did in Honeywill, as he did not take into account the fact that the use of the chemical flashlight, while still dangerous, would not present an unacceptable level of danger if proper precautions were taken (at [75]).

72 In conclusion, Stanley Burton LJ held that while the decision in Honeywill was binding on the English Court of Appeal, it was “so unsatisfactory that its application should be kept as narrow as possible” (at [78]). Drawing particular attention to Slesser LJ’s reference to “a dangerous operation in its intrinsic nature” (see Honeywill at 200, cited at [66] above), he held that the doctrine should be confined only to “activities that are exceptionally dangerous whatever precautions are taken” [emphasis added] (at [78]). Applying that approach to the facts, he held that the activity to be assessed was “welding per se, in the ball mill, not welding in the vicinity of unwetted combustible material” (at [81]). Given the plaintiffs’ concession that welding per se (as opposed to welding in the vicinity of combustible material, which the plaintiffs had submitted was the activity which fell to be assessed: see [69] above) was not ultra-hazardous, the plaintiffs’ appeal was dismissed.

73 The criticisms levelled against the doctrine of ultra-hazardous acts are powerful. It would not be a stretch to say that there are few principles of law which have been the subject of such universal and unstinting criticism, both judicial and academic, at the very highest levels. In Stevens v Brodribb Sawmilling Company Proprietary Limited and another (1986) 160 CLR 16 (“Stevens”), the High Court of Australia held that it had no place in Australian law (at 30 per Mason J; at 43, per Wilson and Dawson JJ). For completeness, I note that while it has been suggested that Burnie had revived the doctrine of ultra-hazardous acts in Australia (see Kit Barker et al, The Law of Torts in Australia (OUP, 5th Ed, 2011) at pp 766–777), the view in the authorities is that Stevens is good law: see Transfield Services (Australia) Pty Ltd v Hall and another appeal [2008] NSWCA 294 at [90]. Mr Han, citing Stevens, put forward a persuasive case for the abolition of the doctrine. ²⁹ For present purposes, it suffices for me to say, like the English Court of Appeal did in Biffa Waste, that I do not consider that it is open to me to take such a bold step. In my judgment, there has been sufficient judicial recognition of its existence which forecloses that option. My task is to determine what the ambit of the doctrine is and to apply it to the facts of this case.

74 The central problem with the doctrine of ultra-hazardous acts, as rightly identified by the English Court of Appeal in Biffa Waste at [73], is that all manner of quotidian activities may be considered to be ultra-hazardous in the right context. Going back to the example of driving, even at the best of times, driving is a dangerous activity, but it cannot sensibly be maintained that it is ultra-hazardous. However, if one speeds through a school zone at 100 km/h resolving all the while never to check for students crossing the road, then the act is doubtlessly ultra-hazardous. How then should the court decide when a person who flags down a taxi should be affixed with a non-delegable duty of care to ensure that the driver exercises reasonable care in driving? The question resolves itself to this: How much of the surrounding circumstances – which may either go towards increasing or decreasing the hazards involved – should be taken into account in determining whether an act is ultra-hazardous?

75 Mr Sreenivasan submitted, contrary to Biffa Waste, that the court should omit consideration of precautionary measures. He contended that to admit consideration of possible precautions “is undesirable because it occasions a high degree of speculation and postulation on the effectiveness of a wide range of possible precautionary measures”. Instead, the court should look only to the “actual degree of hazard created by the activity in its setting.” ³⁰ I rejected this submission for the simple reason, as I pointed out in the preceding paragraph, that it makes no sense to talk about how inherently dangerous an act may be without adverting to the possible steps which may be taken to mitigate the attendant risks.

76 In any event, it seemed to me that the notion of the “hazard created by the activity in its setting” was simply another way of saying that the hazards presented by the activity must be assessed in context. This still begged the question: what is the relevant context? In particular, how does one distinguish between features of the “setting” which go towards the question of breach and those which go towards the character of the act, and therefore affect the nature and content of the duty of care? Consider the example of driving once again. If a taxi driver drives with excessive speed through a school zone, is the fact of his excessive speed a matter which goes to the breach of his duty or is it an issue which goes towards the dangerousness of the act? It is no answer to say that everything turns on what was authorised, for if a negligent act were specifically authorised, we would not need to talk about a non-delegable duty any longer. We would simply be dealing with a case of agency.

77 These are problems which are endemic to the very concept of the doctrine of ultra-hazardous acts. In my judgment, therefore, the best solution is that which was adopted in Biffa Waste, which is to say that in order for an act to be considered ultra-hazardous, it must be “exceptionally dangerous whatever precautions are taken” or, to use the language in Honeywill, it must be “a dangerous operation in its intrinsic nature”. Not only does this approach minimise the difficulties associated with trying to define what surrounding circumstances should be taken into account, it also allows the courts to narrow the application of the doctrine only to that small sliver of cases where it may properly be said to belong.

78 I now turn to the facts of this case. Mr Sreenivasan never argued that the performance of demolition works per se was ultra-hazardous. Instead, his case had always been that it was only demolition considered in the light of (a) the proximity and (b) the relative elevations of the two houses that was ultra-hazardous. ³¹ However, these two points do not go towards showing that the demolition was a dangerous operation in its intrinsic nature. In my judgment, this is sufficient to dispose of this issue.

79 I now turn to the second of Mr Sreenivasan’s arguments. It was not disputed that the respondents owed a duty of care to the appellants to exercise reasonable care in the selection of a competent contractor (see [43] above). The disagreement centred on whether they owed an additional duty, over and above this duty of care in selection, to “ensure that reasonable care was taken [by Esthetix] to avoid harm to the [a]ppellants and to their property.” ³² I will approach this, as the parties did, through an application of the two-stage Spandeck framework: first, I will consider the requirement of legal proximity; second, I will consider whether there are any policy considerations which might operate to militate against the imposition of a duty.

80 The notion of “proximity” is compendious and it has physical, circumstantial, and causal extensions. At its core, it communicates the idea that the defendant and the victim must stand in such a relationship that it may be said that it is proper for the duty of care of a certain type to arise (see Spandeck at [78]–[79]). Mr Sreenivasan urged me to consider the following three factors: (a) the appellants were vulnerable as they had no control over the works; (b) the respondents had “assumed responsibility towards their neighbours” because they had elected to demolish the existing house on their property and to rebuild; and (c) the respondents had control over the performance of the works, as evinced by the fact that they had control of the site, method of work, and could make decisions on workplace safety and health. He submitted that these were cumulatively sufficient to found a duty of care of the sort for which the appellants contended. ³³ I did not agree.

81 What was critical was that the demolition was performed by a third party and the charge was that the respondents were liable for their omission to prevent this third party from causing harm. This is key because, as Lord Goff pointed out in Smith at 270G–271E, one does not generally have a duty of care to prevent third parties from causing damage to others or to their property. Even though that statement was made in the context of deliberate wrongdoing on the part of the third party, the principle which it engages is a general one and applies here: the common law does not generally impose liability for pure omissions. Circumstances in which the law will find that persons have a positive duty to monitor the conduct of third parties are few and far between.

82 In Gary Chan Kok Yew and Lee Pey Woan, The Law of Torts in Singapore (Academy Publishing, 2nd Ed, 2016) at paras 04.043–04.047, the following were given as examples of cases where a positive duty of care to prevent third parties from causing harm to the plaintiff had been found:

(a) In Stansbie v Troman [1948] 2 KB 48, the defendant was a decorator hired by the plaintiff who left the door to the plaintiff’s house unlocked before heading out to purchase some items. Whilst the defendant was away, a thief entered the house and stole several items. The contractual relationship between the parties was held to be sufficient to found a duty of care on the part of the defendant to take reasonable care with regard to the safety of the premises during the performance of his work.

(b) In Home Office v Dorset Yacht Co Ltd [1970] AC 1004, several boys who were under the control and supervision of officers from a Borstal Institution were brought to an island to do some work. In the evening, they escaped and damaged several yachts belonging to the respondent in the process. The majority of the House of Lords held that the combination of (i) special knowledge – the Home Office knew that the Borstal boys had criminal records and had escaped from Borstal Institutions before and were likely to attempt to do so again – and (ii) a special relationship – the Home Office had authority and control over the boys – was sufficient to found a duty of care.

(c) In Haynes v Harwood [1935] 1 KB 146, the defendant had brought a horse-drawn carriage to a public place and left it unattended. A boy threw a stone at the horses, which panicked and bolted. The defendant, a police officer, injured himself while trying to stop the horses. As explained by Lord Goff in Smith at 273A, this was a case in which the defendant had created a source of danger which could foreseeably be triggered by the acts of a third party.

83 The facts of the present situation do not even remotely resemble the facts of these or any other cases in which a positive duty of the sort the appellants contend have been found to exist. I do not think that Mr Sreenivasan had gone any further than to describe the relationship that would ordinarily exist between any neighbours who own adjoining plots of land. This might be enough to demonstrate that the respondents hold a duty of care in respect of anything that they might personally perform, but it cannot be enough to show that they have a duty to ensure the careful performance of works undertaken by any independent contractors they might hire. In my judgment, the present case has none of the incidents of the “necessary relationship between the claimant and defendant” that would justify a finding that the respondents held a positive duty of care to prevent harm from befalling the appellants through Esthetix’s negligence (see Spandeck at [79]).

84 Even if I were wrong about the issue of proximity, I would still have concluded that considerations of policy militated against the finding of a duty of care. I have already touched on some of these points at [35]–[38] above in the context of my discussion of the independent contractor defence. To my mind, there are two main policy objections to a finding of a duty of care here.

85 The first is it would undermine the general principle that persons are not liable for the acts of independent contractors. It was observed in David Tan and Goh Yihan, “The Promise of Universality: The Spandeck Formulation Half a Decade On” (2013) 25 SAcLJ 510 at 541 that the courts would be very slow to find a duty of care if to do so would undermine the coherence of other legal principles. As I have noted throughout this judgment, the principle that employers are not liable for the acts of independent contractors is well-established and it is one around which people have ordered their affairs. A duty will not lightly be found if its effect would be to undermine this principle. The independent contractor doctrine is an outgrowth of an economic necessity. The modern world is simply too complex for businesses to be self-sufficient; businesses need to specialise and they need to hire independent contractors. However, business cannot be carried out if hirers were exposed to liability for the acts of independents contractors over whom they have little control. In the 19th century decision of the House of Lords in David Daniel v The Directors, &C, of the Metropolitan Railway Company (1871) LR 5 HL 45 at 61, Lord Westbury put the point in the following terms:

86 The second policy consideration is that it would expose the respondents and other homeowners in like situation to a potentially indeterminate vista of liability. While Mr Sreenivasan had taken pains to stress that this situation was quite unique, arising as it did from the close proximity of the two homes and their respective elevations, I am quite unable to agree. As the District Judge rightly pointed out at [58] of the GD, Singapore is a country with a high population density and people live very close together. This is a matter of which judicial notice may be taken, and it was also noted in See Toh Siew Kee v Ho Ah Lam Ferrocement (Pte) Ltd and others [2013] 3 SLR 284 (“See Toh Siew Kee”) at [96]. It would follow that cases like the present, far from being rare, can in fact be fairly commonplace. This is further compounded by the fact that it is difficult to discern how such a duty can, as a practical matter, be discharged. The work independent contractors like Esthetix are employed to perform is specialised and hirers do not generally possess the skills to superintend their performance. It would be intolerable if the law were to hold that all landowners who seek to construct homes on their property would have a duty to look continually over the shoulders of the independent contractors they hire to ensure that they take reasonable care in the performance of their tasks. I cannot see how it would be fair, just, and reasonable for such liability (which, as I pointed out above, amounts for all intents and purposes to strict liability) to be imposed in this situation.

87 Last, I would add a brief comment on the issue of insurance. One of the features of this case which received much attention was the fact that the respondents and Esthetix were insured by the same company; however, the insurer had repudiated liability in respect of the respondents but maintained this action on behalf of the respondents. ³⁴ In my judgment, this was neither here nor there. The presence of insurance is neither an incident of the parties’ relationship (and therefore does not go towards the requirement of proximity) nor has it been shown that it is an established feature of all construction contracts (in which event it might plausibly be contended that it is a positive policy consideration in favour of a finding of a duty of care).

88 For these reasons, I dismissed the appeal on the third issue. Before I leave this issue, I make a final point. This concerns the argument, made both by Mr Sreenivasan and by Mr Han, that the Spandeck framework is sufficiently flexible that it may be used to determine the existence of a duty of care in all cases, and that the law need no longer adhere to the old distinctions which continue to pervade the law of non-delegable duties. ³⁵ By analogy, Mr Han cited the case of See Toh Siew Kee where the Court of Appeal subsumed the law on occupiers’ liability within the general law of negligence. Two of the reasons they gave for doing so were: (a) the law on occupiers’ liability was developed before the seminal decision of Donoghue v Stevenson which inaugurated the development of the modern law of negligence; (b) the law on occupiers’ liability was premised on outmoded formal distinctions (eg, the static/dynamic dichotomy) which verged on the arbitrary.

89 While I would not venture so far as to say this in respect of all non-delegable duties, there is a compelling argument in favour of such an approach where ultra-hazardous acts are concerned. The doctrine of ultra-hazardous acts can be assailed on the same two grounds that were raised in See Toh Siew Kee. First, the doctrine became part of the law before the law of negligence was fully developed and, critically, before the requirement of fault became an entrenched part of the law of tort (which happened after Read). The practice of imposing a non-delegable duty (which, as I noted above amounts in practical terms to the imposition of strict liability) by reason of the dangerousness of the act may justly be criticised for being out of step with the modern law of negligence. As Mason J commented in Stevens at 30, “the traditional common law response to the creation of a special danger is not to impose strict liability but to insist on a higher standard of care in the performance of an existing duty.” Second, like the law on occupiers’ liability, the doctrine of ultra-hazardous acts is premised on an unworkable distinction between ultra-hazardous activities and activities which are “merely” dangerous.

90 There may come a day when our Court of Appeal will decide, like the High Court of Australia did in Stevens, that the doctrine no longer has a place in our law and will instead deal with such cases by applying the general principles of the law of negligence (see Stevens at 25). Because of the manner in which the appellants had presented their case (first by addressing the so-called traditional approach and then by applying the general principles of the law of negligence), this issue was considered from both perspectives. However, future courts may no longer have to do both.

91 For the foregoing reasons, I dismissed the appeal and awarded the respondents their costs of the appeal fixed at $10,000 as well as reasonable disbursements.