THE PROMISE OF UNIVERSALITY

• Date: 01 December 2013
• Published date: 01 December 2013
• Author: David TAN LLB (Hons), B Com (Melbourne), LLM (Harvard), PhD (Melbourne); Associate Professor, Faculty of Law, National University of Singapore. GOH Yihan LLB (Hons) (National University of Singapore), LLM (Harvard); Advocate and Solicitor (Singapore); Assistant Professor, Faculty of Law, National University of Singapore.

The landmark decision by the Court of Appeal in Spandeck Engineering (S) Pte Ltd v Defence Science & Technology Agency[2007] 4 SLR(R) 100 undoubtedly broke new ground in Singapore by unifying inconsistent case law into a single, universal test for determining a duty of care in the law of negligence. This article surveys how the Spandeck formulation has been applied by the Court of Appeal in the last half decade, and evaluates its effectiveness in the determination of duty in a wide range of scenarios involving different types of harm. It aims to provide a set of practical guidelines that will benefit the legal community in advising clients, drafting pleadings, framing submissions for court, and even drafting of judgments.

I. Introduction

1 In 2007, the Singapore Court of Appeal declared in Spandeck Engineering (S) Pte Ltd v Defence Science & Technology Agency (“Spandeck”)1“a single test … to determine the imposition of a duty of care in all claims arising out of negligence, irrespective of the type of the damages claimed”.2 The preference for this two-stage test set aside other tests, such as those based on assumption of responsibility,3 pure policy4 or incrementalism. The court has since adhered to its bold claim of

universality; a year later in Ngiam Kong Seng v Lim Chiew Hock (“Ngiam”),5 it reiterated “the ideal envisioned in Spandeck of having ‘a single test’” [citation omitted]6 but conceded that “in determining whether the requisite proximity is present in a particular case, much will turn on the precise factual matrix concerned”.7 The court explained there that “there must, in principle as well as in logic, justice and fairness, be a holistic and integrated analysis of the relevant factual matrix both from the perspective of proximity (as between the parties) and from the perspective of public policy (on a broader societal level)”.8 The elegant statement of the Spandeck formulation belies its doctrinal complexity.9 Indeed, the search for a definitive test to determine whether a duty of care should be imposed for negligence liability has plagued common law jurisdictions for over half a century. From a two-stage test in Anns v Merton London Borough Council (“Anns”)10 to a three-part test in Caparo Industries plc v Dickman (“Caparo”)11 to a multi-factorial approach in Graham Barclay Oysters v Ryan (“Graham Barclay Oysters”),12 courts have experimented with different formulations with varying degrees of success. The disparate types of harm that a claimant may suffer, coupled with the myriad factual scenarios in which the harm may arise, present significant challenges for any test that claims to be capable of being universally applied to all claims arising out of negligence.

2 This article surveys how the Spandeck formulation has been applied by the Singapore courts in the last half decade, and evaluates its effectiveness in the determination of the existence of a duty of care in a wide range of scenarios involving different types of harm.13 It ultimately endeavours to provide a set of practical guidelines that will benefit “the daily business of advising clients, drafting pleadings, framing submissions

for court, and even drafting of judgments”.14 In this light, Part II considers the Spandeck formulation's claim of universality and the challenges of applying a “single test” to different types of harm. Drawing from the identified limitations of universality, it postulates how particular types of harm may influence the stage-wise application of the Spandeck test. Parts III, IV and V then discuss the role and the merits of the Spandeck requirements of factual foreseeability, proximity and public policy respectively, and evaluate how the Singapore courts have applied the Spandeck formulation to different factual scenarios.

II. Examining the universality claim

3 The Spandeck formulation has been referred to as a “universal” test. To understand this, it is necessary to differentiate between degrees of universality. More particularly, the Spandeck claim of universality is to be understood as a universal framework, and not as a universal test to be applied identically in every case. The type of harm suffered will therefore affect the factors relevant to each stage of the Spandeck formulation, without affecting the claim of universality.

A. Different degrees of “universality”

4 The decided cases reveal two definitions of universality. The broad definition posits that the same test applies in every fact situation, independent of nuanced considerations of particular facts. A narrow definition provides that the same framework always applies but this does not preclude consideration of factors that may not be universally applicable. This narrower sense of universality can be seen when the premise of universality is contrasted with an approach based on specific categories of people or situations.15 Before Heaven v Pender16 was decided, the existing liability was decided on the basis of fitting the relevant factual scenario into a pre-existing category where a duty of care had been held to exist.

5 The first hint of narrow universality came in Heaven v Pender,17 where Brett MR, in a minority judgment, purported to find a general principle.18 Brett MR's statement represented a departure from the category-based approach by formulating a general principle, which,

if satisfied, would result in liability. In Donoghue v Stevenson,19 this statement was to influence Lord Atkin's formulation of the “neighbour principle”. While Lord Atkin thought Brett MR's statement too wide,20 he did not disagree with the claim of generality embodied within. In fact, the main difference between the “neighbour principle” and Brett MR's statement is the addition of proximity to the statement.21 The premise of narrow universality can be seen further in Home Office v Dorset Yacht Co Ltd,22 where Lord Diplock made clear that both inductive and deductive reasoning were to be used in establishing duty. Inductive reasoning involves identifying the relevant characteristics common to the kinds of conduct and relationship between the parties, whereas deductive reasoning involves assessing whether, in all cases where the conduct and relationship possesses certain characteristics, a duty of care arises.23 However, neither form of reasoning elevates the claim of universality to an immutable test. In fact, both mandate the need to consider past cases relative to the set of facts before the court. The claim for universality is thus a narrower one in that what is universal is only the general framework, not the test itself.

6 The broader sense of universality is usually thought to originate from Anns, where, amongst other things, Lord Wilberforce said that “in order to establish that a duty of care arises in a particular situation, it is not necessary to bring the facts of that situation within those of previous situations in which a duty of care has been held to exist” [emphasis added].24 This has been taken to mean that there was a “universal test which was applicable without resort to prior decisions”.25 This broader claim of universality forsakes all references to previous cases; the universality transcends being a mere framework to an independent and exhaustive test.

7 It is important to distinguish these two notions of universality. Later decisions cast doubt on the Anns test for suggesting too broad a sense of universality.26 However, the error in contemporary analysis, exemplified by the House of Lords decision in Caparo,27 is not distinguishing between these two versions of universality. The retreat from Anns went further than it needed to be. In Caparo, for example, Lord Bridge purported to prefer an “incremental approach”; to developing new categories of negligence via analogy to the establish

categories instead of the general and broad formulation in Anns.28 However, earlier in his speech, Lord Bridge had laid down a three-part test that seems to hint at some universality, as opposed to his call for a categorisation approach, which is antithetical to such universality. In the same case, Lord Roskill likewise preferred the “traditional categorisation of cases” as compared to the “somewhat wide generalisations that leave their practical application matters of difficulty and uncertainty”.29 These statements regard universality as only referring to the broader axiom while ignoring its narrower sense. However, the better view is that it is the test that is universal, but its application has always been particular to the facts.

8 Spandeck adopted the narrower concept of universality: the Spandeck test is to be regarded as an umbrella framework to be applied in every case, but its actual application can admit of varying considerations. Incrementalism is then employed as a methological aid within the Spandeck formulation. Indeed, on its own, incrementalism “lacks any firm legal basis but appears to reflect … the general ‘common law’ approach of drawing analogy from prior precedents”.30 However, to speak of drawing analogies raises a separate question: analogy with what? It would be wrong to draw an analogy on a composite basis with decided cases; that is, if similar facts give rise to a duty, then the current case should also give rise to duty. This nullifies the reality that the Spandeck test is a multi-stage test that comprises distinct elements, each to be decided separately. Incrementalism functions at each specific stage, and it entails considering different factors that inform whether that particular element has been satisfied. There should not be a presumption of prima facie duty simply on the basis of general similarity in facts alone. Each stage has to be considered separately with no pre-existing notion of duty.

B. The type of harm

9 The type of harm is important in adhering to incrementalism within the Spandeck formulation even though a “single test” is to be applied regardless of the type of harm. The latter caveat simply...