R obinson v. C hief Constable of West Yorkshire Police: Taking Duty Back to Basics

Abstract

In the United Kingdom, as in some other parts of the Commonwealth, the courts have long embarked on a search for a general formula which can be applied to determine whether a duty of care arises in any given negligence case. In 2018, the United Kingdom Supreme Court delivered its judgment in Robinson v. Chief Constable of West Yorkshire Police , in which a plurality strikingly rejected as misconceived attempts to identify such a general test. Duty cases, the Supreme Court held, are to be decided by reference to precedent where applicable, or (where no such precedent applies) by analogy with the existing authorities. This approach curtails the role of policy in the duty of care enquiry. In this article, the decision in Robinson is put in context and its significance explained. It is argued that the approach to the duty question adopted in Robinson should be welcomed. Not only is that approach the best of the alternatives available, but it is the only one that is consistent with the methodology of the common law and the rule of law.

Full text

James Goudkamp* and Donal Nolan
Robinson v. Chief Constable of West Yorkshire
Police: Taking Duty Back to Basics
https://doi.org/10.1515/jtl-2023-0029
Received July 5, 2023; accepted July 5, 2023; published online July 25, 2023
Abstract: In the United Kingdom, as in some other parts of the Commonwealth, the
courts have long embarked on a search for a general formula which can be applied to
determine whether a duty of care arises in any given negligence case. In 2018, the
United Kingdom Supreme Court delivered its judgment in Robinson v. Chief Constable
of West Yorkshire Police, in which a plurality strikingly rejected as misconceived
attempts to identify such a general test. Duty cases, the Supreme Court held, are to be
decided by reference to precedent where applicable, or (where no such precedent
applies) by analogy with the existing authorities. This approach curtails the role of
policy in the duty of care enquiry. In this article, the decision in Robinson is put in
context and its significance explained. It is argued that the approach to the duty
question adopted in Robinson should be welcomed. Not only is that approach the best
of the alternatives available, but it is the only one that is consistent with the meth-
odology of the common law and the rule of law.
Keywords: duty of care; negligence; Robinson v. Chief Constable of West Yorkshire
Police
1 Introduction
Few legal issues have excited as much attention, academic and judicial, as the duty of
care element of the tort of negligence. Leading duty cases have long formed the
foundation of torts textbooks, and a steady stream of judgments regarding the duty
concept has flowed from ultimate appellate courts in the United Kingdom and
elsewhere in the Commonwealth. It is unsurprising that the duty element has
garnered so much attention. In earlier times, before the demise of the civil jury in
most of the common law world, it played a vital role in maintaining judicial control
over negligence claims. In the modern era, the duty concept enables courts to lay
*Corresponding author: James Goudkamp, Professor of the Law of Obligations, University of Oxford;
Fellow and Tutor in Law, Keble College, Oxford, UK, E-mail: james.goudkamp@law.ox.ac.uk
Donal Nolan, Professor of Private Law, University of Oxford; Francis Reynolds and Clarendon Fellow and
Tutor in Law, Worcester College, Oxford, UK
J. Tort Law 2023; 16(1): 125–142
Open Access. © 2023 the author(s), published by De Gruyter. This work is licensed under the
Creative Commons Attribution 4.0 International License.

down categorical rules regarding the reach of negligence liability as a matter of law.
1
Because the existence or otherwise of a duty of care is a question of law, the duty
element is often used as vehicle for disposing of claims without a trial via the
summary judgment procedure or facility to strike out the claim on the ground that it
discloses no cause of action.
It is probably fair to say that the duty concept is a barometer for judges’senti-
ments regarding tort law. Duty cases often provide an indication as to whether
particular judges would prefer tort law to play an enlarged or reduced role in society.
They also offer a window into the function that judges discharge, and perceive
themselves as discharging, in the tort process. They sometimes reveal the courts to be
politically active institutions. On other occasions, they show judges remaining (or
appearing to remain) aloof from the types of policy debates that typify the legislative
process. The recent decision of the United Kingdom Supreme Court in Robinson v.
Chief Constable of West Yorkshire Police,
2
on which this article focuses, bears all of
these points out. As we will see, it involves a dramatic return to the conventional
methodology of the common law.
Before turning to Robinson itself, it is important to provide some context
regarding the English law concerning the duty of care concept. In 1932, the House of
Lords delivered judgment in the famous case of Donoghue v. Stevenson.
3
In his
speech, Lord Atkin articulated the idea that one owes a duty of care to one’s
“neighbours”.
4
This led him to hold that a manufacturer of ginger beer owed a duty of
care to a consumer who was allegedly injured by ingesting the remains of a snail
which had been sealed within one of its bottles. The fact that the consumer and
manufacturer were not in privity of contract, because the bottle had been sold via a
retailer and to a friend of the consumer, did not preclude this conclusion.
5
Gradually, this “neighbour principle”fuelled the belief that it was possible and,
indeed, desirable to formulate a general test for the existence of a duty of care, that is
1See Donal Nolan, Deconstructing the Duty of Care, 129 Law Q. Rev 559, 567–573 (2013).
2[2018] UKSC 4, [2018] AC 736.
3[1932] AC 562 (HL).
4Id., 580 (“acts or omissions which any moral code would censure cannot in a practical world be
treated so as to give a right to every person injured by them to demand relief. In this way rules of law
arise which limit the range of complainants and the extent of their remedy. The rule that you are to
love your neighbour becomes in law, you must not injure your neighbour; and the lawyer’s question,
Who is my neighbour? receives a restricted reply. You must take reasonable care to avoid acts or
omissions which you can reasonably foresee would be likely to injure your neighbour. Who, then, in
law is my neighbour? The answer seems to be –persons who are so closely and directly affected by
my act that I ought reasonably to have them in contemplation as being so affected when I am
directing my mind to the acts or omissions which are called in question”).
5The sister case in the United States is, of course, MacPherson v. Buick Motor Co., 111 NE 1050 (NY
1916).
126 J. Goudkamp and D. Nolan

to say, an equation that could be applied in all types of negligence cases to produce an
answer to the duty issue. Thus, in 1970 Lord Reid said that the time had come “when
we can and should say that [the ‘neighbour principle’] ought to apply unless there is
some justification or valid explanation for its exclusion”.
6
In 1978, in Anns v. Merton
London Borough Council, Lord Wilberforce drew upon Lord Reid’sdictum in laying
down a general two-stage test for the determination of the duty issue.
7
This marked
the high point of what has sometimes been dubbed the “modern approach”to the
duty enquiry. The first step of the Anns test involved asking whether injury to a
person such as the claimant was a reasonably foreseeable result of a failure by the
defendant to exercise reasonable care. If that stage was satisfied, a duty would
ordinarily arise unless the defendant was able to show by reference to policy factors
that a duty should not be imposed.
Eventually, concerns that the Anns test involved an overly expansive approach
to the duty concept and had led to unpredictable outcomes resulted in its being
replaced. For the better part of three decades, the UK courts frequently applied an
alternative, more conservative, general duty test. This was the so-called “Caparo test”
(or “three-stage test”), which was derived from the speech of Lord Bridge in Caparo
Industries plc v. Dickman,
8
a case decided in 1990. Pursuant to that test a duty of care
would arise where (1) injury to a person such as the claimant was reasonably fore-
seeable, (2) the parties were in a relationship of “proximity”and (3) imposing a duty
would be “fair, just and reasonable”. The Caparo test, like its predecessor, was a
policy-sensitive test, owing in large part to its third and final stage. Like the Anns test,
it too was treated (by those who employed it) as a universal solution to the duty issue.
There matters stood until the litigation in Robinson. As we shall see, the Supreme
Court characterised duty cases that had been decided over the course of the pre-
ceding decades as having been largely misunderstood and rejected the suggestion
that the duty enquiry could be dealt with by way of a general test. Relatedly, the
Supreme Court left a far smaller role for the deployment of policy by judges in duty
cases and placed heavy emphasis on the importance of principle and precedent. On
any measure, therefore, Robinson is a canonical case, and one of the most significant
that the Supreme Court has decided in its relatively brief history (having replaced the
House of Lords as the apex court in the United Kingdom in 2009) in any branch of the
law.
In addition to outlining the way in which the UK courts had, pre-Robinson,
searched for a general test for the existence of a duty of care, it is important that we
explain a few other matters by way of background. One vital point is that in cases
6Home Office v. Dorset Yacht Co [1970] AC 1004 (HL) 1027. See also at 1054 (Lord Pearson).
7[1978] AC 728 (HL) 751–752.
8[1990] 2 AC 605 (HL) 617–618.
Robinson v. Chief Constable of West Yorkshire 127

involving foreseeable physical damage caused by a positive act, the defendant will be
hard pressed to sustain any argument that no duty was owed.
9
It is only in rare
circumstances, frequently where a judge has gone (or been led) astray, that any duty
issue arises in such cases.
10
In this regard, it appears that the relevant English law
corresponds with that in jurisdictions in the United States.
11
Conversely, there are well-established no-duty categories, all of which are sub-
ject to one or more exceptions. The main such categories are as follows. First, the UK
courts have established a bright-line rule to the effect that defendants do not
generally owe a duty of care to guard against the risk of causing the claimant to suffer
pure economic loss. The primary exception to this rule is where the defendant has
assumed responsibility to take reasonable care to protect the claimant from suffering
such loss.
12
Second, there is usually no duty of care owed by defendants who have
merely failed to act (or, as the UK courts now tend to put it, “failed to confer a benefit”
on the claimant). This no-duty rule is subject to a series of exceptions,
13
including
where the defendant has assumed a positive duty to look out for the claimant’s
interests. Third, a defendant who negligently kills, injures or imperils another
frequently owes no duty of care to a third party who suffers mental harm as a
result.
14
9See JAMES GOUDKAMP &DONAL NOLAN,WINFIELD &JOLOWICZ ON TORT §5.014 (20th ed. 2020).
10 Examples include Sam v. Atkins [2005] EWCA Civ 1452; Darnley v. Croydon Health Services NHS
Trust [2017] EWCA Civ 151, [2018] QB 783 (reversed on appeal: [2018] UKSC 50, [2019] AC 831).
11 See JOHN C.P. GOLDBERG &BENJAMIN C. ZIPURSKY,THE OXFORD INTRODUCTIONS TO U.S. LAW:TORTS 78 (2010)
(“Speaking generally, negligence suits arising out of acts by a defendant that have caused physical
injuries and property damage to a plaintiff—or at least have done so without the intervention of
other actors’wrongful acts—tend not to raise difficult duty questions. This is because courts have
fashioned the relevant duty rules quite broadly. For example, it is common for courts to say that any
person who engages in a course of conduct—whether walking, driving, operating a machine, or
providing a service—owes a duty to exercise due care against physically harming anyone whom he
might reasonably foresee physically harming were he to perform that conduct carelessly”). See also §7
of the Restatement (Third) of the Law of Torts: Liability for Physical and Emotional Harm (2010), which
provides “(a) An actor ordinarily has a duty to exercise reasonable care when the actor’s conduct
creates a risk of physical harm. (b) In exceptional cases, when an articulated countervailing principle
or policy warrants denying or limiting liability in a particular class of cases, a court may decide that
the defendant has no duty or that the ordinary duty of reasonable care requires modification”.
12 See Hedley Byrne & Co Ltd v. Heller & Partners Ltd [1964] AC (HL). Regarding the assumption of
responsibility concept, see Donal Nolan, Assumption of Responsibility: Four Questions, 72 C.L.P. 123
(2019).
13 For prominent treatments, see Smith v. Littlewoods Organisation Ltd [1987] AC 241 (HL) 272–274
(dealing with exceptions to the closely allied principle that there is ordinarily no duty to take steps to
control third parties); Stovin v. Wise [1996] AC 923 (HL) 944.
14 See Alcock v. Chief Constable of South Yorkshire Police [1992] 1 AC 310 (HL).
128 J. Goudkamp and D. Nolan

2 The Facts of Robinson and Decisions Below
One afternoon in 2008, a police officer witnessed an apparent drug deal in a park in
Huddersfield, a town in the north of England. He thereafter shadowed the suspect,
who entered a bookmakers. The officer decided not to make an arrest at this stage
partly due to a concern that doing so would endanger the other customers. By the
time that the suspect had left the bookmakers, more police officers had arrived on the
scene and the decision was made to arrest him in the street. However, the suspect
resisted and in the ensuing scufflea“relatively frail”76-year-old pedestrian was
knocked over and injured. She sued the police for negligence.
The trial judge found that the officers concerned had carried out the arrest
carelessly. Nevertheless, he held that the police were not liable in view of the decision
of the House of Lords in Hill v. Chief Constable of West Yorkshire.
15
He said that case
conferred on the police an immunity from suit in negligence. In Hill, the House had
held that the police did not owe a duty of care to a woman who had been murdered by
a notorious serial killer. Although Lord Keith had focused in his speech in Hill mainly
on what he considered to be the lack of “proximity”between the police and the
murder victim, he also relied on several policy considerations which he considered
militated against recognition of a duty of care. These included a concern that
imposing liability might result in police work being carried out “with a detrimentally
defensive frame of mind”.
16
The Court of Appeal dismissed an appeal by the claimant in Robinson. Hallett LJ,
delivering a judgment with which the other members of the Court of Appeal agreed,
applied the three-stage test derived from Caparo. The third stage of the test meant,
Hallett LJ said, that a court would “only impose a duty where it considers it right to do
so on the facts”.
17
It would rarely be right to do so, she remarked, where the police
were sued for negligence in respect of their conduct “in the course of investigating
and supressing crime and apprehending offenders”, since the interests of the public
would not be best served by imposing a duty of care in such cases.
18
Hallett LJ also
argued that “provided the police act within reason, the public would prefer to see
them doing their job and taking drug dealers offthe street”. Any risk thereby posed to
passers-by such as the claimant was “trumped by the risk to society as a whole”.
19
Hallett LJ additionally considered that the claimant had been injured by the suspect
15 [1989] AC 53 (HL).
16 Id., 63 (Lord Keith).
17 Robinson v. Chief Constable of West Yorkshire Police [2014] EWCA Civ 15, [2014] P.I.Q.R. P14 [40].
18 Id., [46].
19 Id., [47].
Robinson v. Chief Constable of West Yorkshire 129

and that the police had simply failed to protect her from him.
20
Finally, Hallett LJ held
that the police officers had not acted negligently.
This short summary of HallettLJ’s judgment does not, of course, do full justice to her
analysis. Nevertheless, it is instructive to draw attention to three features of her
reasoning. The firstisthatHallettLJinvokedtheCaparo test, and particularly the third
stage thereof, to justify what appears to be an untrammelled discretion whereby a court
(seemingly at any level of the hierarchy) can refuse to recognise a duty of care on the
ground that it considers it “righttodoso”on the facts of the individual case. That
approach seems to leave minimal or no room for the doctrine of precedent. A judge can
simply hold that no duty of care is owed even if another court in a similar case has
reached the contrary conclusion. The second is that Hallett LJ appealed to public
sentiment. She referred to what the public would prefer, which is an empirical claim for
which there was apparently no empirical evidence (certainly, Hallett LJ cited none). And
the third is that Hallett LJ characterised the conduct of the police –a decision to carry out
an arrest on which they then acted –as a mere failure to protect the claimant frombeing
injured by the suspect. She reached that conclusion despite the fact that, independently
of the attempt to arrest him, the suspect posed no danger to the claimant.
Pausing here, it is obvious that something went seriously wrong with the Court
of Appeal’s decision. A law-abiding senior citizen who had been negligently injured
by positive conduct on the part of the police was told that no liability could arise
because, in the Court of Appeal’s view, imposing such liability would have detri-
mental consequences. She was also told that the court of public opinion would (if
asked, though it had not been) take the same view. None of this is consistent with the
general principle, to which we have referred above, that a duty of care ordinarily
arises in cases in which foreseeable physical damage is caused by a positive act. Here,
not only was the very risk that materialised foreseeable but it was actually foreseen.
These are far from the only difficulties with the Court of Appeal’s decision. But it is
clear, even at this stage and for these brief reasons, that it was deeply flawed.
3 An Intervening Development
The claimant appealed to the Supreme Court. While her appeal was pending, the
Supreme Court decided another police negligence case of great significance, namely
Michael v. Chief Constable of South Wales Police.
21
The decision in Robinson cannot be
20 Id., [53].
21 [2015] UKSC 2, [2015] AC 1732. See further James Goudkamp, A Revolution in Duty of Care?, 131 LAW
Q. REV. 519 (2015); Sandy Steel, State Liability for Failure to Save a Person from Being Killed,J.TORT LAW
(forthcoming).
130 J. Goudkamp and D. Nolan

properly understood without our saying a few words about that case. In Michael,it
was held by a majority of the Supreme Court that the police did not owe a duty of care
to a woman who made an emergency call reporting a threat to kill her, and who was
then murdered before the police arrived at her house. Although the facts of the two
cases were very different, the reasoning in Michael cast significant doubt on the
Court of Appeal’s decision in Robinson, for three main reasons.
First, Michael poured cold water on the idea that duty of care questions could be
resolved by the application of a general test. In his majority judgment, Lord Toulson
said:
From time to time the courts have looked for some universal formula or yardstick [for deter-
mining duty of care issues], but the quest has been elusive. And from time to time a court has
used an expression in explaining its reasons for reaching a particular decision which has then
been squashed and squeezed in other cases where it does not fit so aptly.
22
Lord Toulson then contended that there was no such thing as a Caparo test. He said
that “[p]aradoxically, [the relevant] passage in Lord Bridge’s speech [in Caparo] has
sometimes come to be treated as a blueprint for deciding cases, despite the pains
which the author took to make clear that it was not intended to be any such thing”.
23
By explicitly repudiating the use of the Caparo test, Lord Toulson had undermined
the main basis for the Court of Appeal’s decision in Robinson.
The second reason has to do with the fact that the decision in Michael was
premised on the logic that public defendants and private defendants were essentially
subject to the same rules as far as the law of negligence was concerned.
24
In Michael,
this “equality principle”had favoured the defendants. That was because, as the
Supreme Court emphasised, just as a private individual would not have been legally
obliged to answer the woman’s cry for help, nor did any liability attach to the police
for not doing so expeditiously. However, on the facts of Robinson this equality
principle favoured the claimant, since a private individual who carelessly collided
with a passer-by in the street would ordinarily be subject to negligence liability for
any injury that resulted.
Thirdly, while both the trial judge and Court of Appeal in Robinson had
considered that the police enjoyed a broad immunity from negligence liability
pursuant to Hill,inMichael Lord Toulson had expressly denied that Hill supported
any such idea. On the contrary, he said that Hill had “recognised that the general law
22 [2015] UKSC 2, [2015] AC 1732, [102]–[103].
23 Id., [106].
24 Id., [101] (remarking that the general principles regarding omissions “have been worked out for
the most part in cases involving private litigants, but they are equally applicable where D is a public
body”).
Robinson v. Chief Constable of West Yorkshire 131

of tort applies as much to the police as to anyone else”,
25
and that although the House of
Lords had admittedly used the language of “immunity”, this turn of phrase had been “not
only unnecessary but unfortunate”.
26
Thiswassaidtobebecausethecourts’refusal to
imposeadutyonthepolicetorescuevictimsofcrimeflowed not from any immunity but
from the familiar rule that there is ordinarily no liability for omissions.
27
4 The Supreme Court’s Decision
Unsurprisingly given the decision in Michael, the Supreme Court unanimously allowed
the claimant’s appeal in Robinson and held that the police had owed her a duty of care.
Delivering the main judgment, Lord Reed (who had been one of the judges who heard the
appeal in Michael, and who had concurred with Lord Toulson’s reasons) said that the case
involved positive action by the police rather than, as the Court of Appeal had thought, an
omission. He observed that the claimant’sargumentwas“notthatthepoliceofficers failed
to protect her against the risk of being injured, but that their actions resulted in her being
injured”.
28
This is clearly correct. A good litmus test for determining whether a case
involves an omission is to ask whether, if the defendant were removed from the situation
altogether, the claimant would still have been injured. Plainly, that test was not satisfied in
Robinson:hadtheofficersnotsoughttoarrestthesuspect,thetusslewouldnothave
happened and the claimant would not have been hurt. Since Robinson was a case of
positive action causing foreseeable physical injury, that was sufficient, applying orthodox
negligence principles, to establish that the officers had owed the claimant a duty to take
reasonable care to guard against the risk of causing her to suffer physical injury. As for the
breach of duty requirement, Lord Reed, agreeing with the trial judge and disagreeing with
the Court of Appeal, concluded that it was satisfied on the basis that the officers should
have noticed the claimant and delayed the arrest until she was out of harm’s way.
29
In addition to correcting the Court of Appeal’s error in holding that the omissions
doctrine was engaged when it was not, Lord Reed seized the opportunity that the
appeal presented to restate the proper approach to the determination of whether a
duty of care was owed. He distinguished between those categories of relationship in
which the courts had already decided a duty of care did or did not arise and those in
which no such determination had been made. With regard to the former, he said that
“there are many situations in which it has been clearly established that a duty of care
25 Id., [37].
26 Id., [44].
27 Id., [115]–[116].
28 Id., [73]. See also at [82] (Lord Mance) and [122] (Lord Hughes).
29 Id., [75]–[78].
132 J. Goudkamp and D. Nolan

is or is not owed”, giving as examples the duties of care owed by “motorists to other
road users, by manufacturers to consumers, by employers to their employees, and by
doctors to their patients”.
30
In such cases, once the decision has been made that a
duty of care is owed, then that decision, Lord Reed made clear, will apply to all future
cases of the same kind. Furthermore, since in cases of this type “a consideration of
justice and reasonableness forms part of the basis on which the law has arrived at the
relevant principles”it is “unnecessary and inappropriate to reconsider whether the
existence of the duty is fair, just and reasonable”, unless, that is, the court has been
invited to depart from an established precedent.
31
As for cases falling within the latter category (“novel type[s] of case”), Lord Reed said:
[T]he characteristic approach of the common law in such situations is to develop incrementally
and by analogy with established authority. The drawing of an analogy depends on identifying
the legally significant features of the situations with which the earlier authorities were con-
cerned. The courts also have to exercise judgement when deciding whether a duty of care should
be recognised in a novel type of case.
32
In view of the foregoing, Lord Reed described the overall approach that the courts
should take in deciding whether or not a duty of care was owed as follows:
In the ordinary run of cases, courts consider what has been decided previously and follow the
precedents (unless it is necessary to consider whether the precedents should be departed from).
In cases where the question whether a duty of care arises has not previously been decided, the
courts will consider the closest analogies in the existing law, with a view to maintaining the
coherence of the law and the avoidance of inappropriate distinctions. They will also weigh up
the reasons for and against imposing liability, in order to decide whether the existence of a duty
of care would be just and reasonable.
33
Consistently with the foregoing, Lord Reed was clear that the supposed Caparo test
did not exist. He said:
The proposition that there is a Caparo test which applies to all claims in the modern law of
negligence, and that in consequence the court will only impose a duty of care where it considers
it fair, just and reasonable to do so on the particular facts, is mistaken. As Lord Toulson pointed
out in his landmark judgment in [Michael], that understanding of the case mistakes the whole
point of Caparo, which was to repudiate the idea that there is a single test which can be applied
in all cases in order to determine whether a duty of care exists, and instead to adopt an approach
based, in the manner characteristic of the common law, on precedent, and on the development
of the law incrementally and by analogy with established authorities.
34
30 Id., [26].
31 Id.
32 Id., [27].
33 Id., [29].
34 Id., [21].
Robinson v. Chief Constable of West Yorkshire 133

It is important to emphasise that Lord Reed did not merely deny the existence of any
role for the Caparo test but disapproved of the very idea that the House of Lords in
Caparo had intended to lay down a “test”at all. Like Lord Toulson in Michael, Lord
Reed stressed that in Caparo Lord Bridge had expressly denied that a single general
principle could provide a practical test to determine whether a duty of care was
owed, and considered it “ironic that …Lord Bridge’s speech has been treated as
laying down such a test”.
35
On the contrary, Lord Bridge had adopted “an incremental
approach, based on the use of established authorities to provide guidance as to how
novel questions should be decided”, and “it was that approach, and not a supposed
tripartite test”which Lord Bridge had then used to decide the case.
36
Incremental
reasoning, by analogy from existing authority, was the true Caparo approach, and
“the existence of a duty of care did not depend on the application of a ‘Caparo test’to
the facts of a particular case”.
37
5 Policy
In this section and those that follow, we explore several issues of general importance
that Robinson raises. The first, which we address under this heading, concerns the
place of policy in the duty analysis. Hill had widely been interpreted as laying down a
policy-based immunity for the police in negligence cases concerned with the pre-
vention and investigation of crime (and had been so interpreted by the trial judge
and Court of Appeal in Robinson). However, Lord Reed instead explained that case (as
had Lord Toulson in Michael) by reference to the omissions doctrine. He summed up
his discussion in this regard as follows:
[T]here is no general rule that the police are not under any duty of care when discharging their
function of preventing and investigating crime. They generally owe a duty of care when such a
duty arises under ordinary principles of the law of negligence, unless statute or the common law
provides otherwise.
38
Lord Reed also made some more general observations about the role of policy in the
negligence context and the duty element in particular. He said:
[I]t is important to understand that [discussions of policy considerations] are not a routine
aspect of deciding cases in the law of negligence, and are unnecessary when existing principles
provide a clear basis for the decision, as in the present appeal …The absence of a duty towards
35 Id., [24].
36 Id., [25].
37 Id., [30].
38 Id., [70].
134 J. Goudkamp and D. Nolan

victims of crime, for example, does not depend merely on a policy devised by a recent gener-
ation of judges in relation to policing: it is based on the application of a general and long-
established principle that the common law imposes no liability to protect persons against harm
caused by third parties, in the absence of a recognised exception such as a voluntary assumption
of responsibility.
39
At the same time, Lord Reed made it clear that he was not saying that policy concerns
had no place in the duty of care enquiry: on the contrary, while the courts were not
“policy-making bodies in the sense in which that can be said of”law reform bodies or
government departments, he accepted that “the exercise of judgement about the
potential consequences of a decision”had a part to play when a court was asked to
decide whether a novel duty of care existed,
40
at least where “established principles”
did not provide a clear answer to the duty question.
41
It follows from this discussion
that although policy considerations do have a role to play in the duty enquiry, this is
limited to novel cases. Further, the role of policy arguments in novel cases is sec-
ondary to the role of “established principles”.
It was this aspect of Lord Reed’s analysis that provoked the strongest reaction
from some of the other judges who heard the appeal, albeit that they all agreed on the
outcome. According to Lord Mance, it would be “unrealistic to suggest that, when
recognising and developing an established category [of negligence liability], the
courts are not influenced by policy considerations”, and the reality was that “in
recognising the existence of any generalised duty in particular circumstances they
are making policy choices”.
42
Similarly, Lord Hughes said that earlier judicial reli-
ance on policy considerations was “simply too considered, too powerful and too
authoritative in law to be consigned to history”, and that it was not possible to treat
such considerations as “no more than supporting arguments”.
43
When considering the tensions between the Justices’approaches to the role of
policy, it is important to distinguish between the “is”question of the extent to which
the courts have in fact relied on policy considerations to decide duty cases and the
“ought”question of the extent to which it is appropriate for them to base duty
decisions on such considerations. In relation to the “is”question, empirical research
by Plunkett regarding the use of policy-based reasoning in the UK shows that only in
a small minority of duty cases at the ultimate appellate level in which reliance was
placed upon policy considerations were they determinative of the outcome of the
duty enquiry.
44
Perhaps the best example of merely incidental reliance on policy is
39 Id., [69].
40 Id.
41 Id., [42].
42 Id., [84].
43 Id., [113].
44 See JAMES PLUNKETT,THE DUTY OF CARE IN NEGLIGENCE 203 (Table 13) (2018).
Robinson v. Chief Constable of West Yorkshire 135

Hill itself, which is easily explicable (and was explained in Robinson) by reference to
the omissions doctrine.
However, Plunkett’s research also accords with Lord Mance’s observation in
Robinson that there are undoubtedly some earlier cases decided at the highest level
where it was accepted that a duty of care could be denied on policy grounds. A good
example of such a decision is Smith v. Ministry of Defence,
45
in which the House of
Lords upheld the existence of a “combat immunity”applicable to anything done in
the course of military operations against an enemy, an immunity justified on the
ground that the existence of a duty of care might hamper the conduct of the military
in wartime. This immunity means that even in cases in which a defendant combatant
causes foreseeable physical injury or physical damage to property by their positive
conduct (and where one would therefore expect there to be a duty of care in
accordance with the general principles upheld in Robinson), no liability in negligence
can arise.
As for the “ought”question, addressing it would take us deep into the theory of
tort law as well as jurisprudential works concerning the function of judges. Space
precludes our embarking on an analysis in this regard. It must suffice for present
purposes to observe that Lord Reed’s position that it is legitimate for judges in novel
cases to weigh policy considerations but that such considerations are of secondary
significance, accords with much academic thinking, especially in the Common-
wealth.
46
At the same time, however, there are policy maximalists, who view with
scepticism attempts to distinguish between policy and principle,
47
and policy mini-
malists who consider any recourse to policy to be either inconsistent with the nature
of private law or the business of judging or both.
48
45 [2013] UKSC 41, [2014] AC 52.
46 See, e.g., Dilan A. Esper & Gregory C. Keating, “Putting Duty in Its Place”: A Reply to Professors
Goldberg and Zipursky,41L
OY. L.A. L. REV. 1225, 1247–1249 (2008); STEPHEN PERRY,The Role of Duty of Care
in a Rights-Based Theory of Negligence Law,inT
HE GOALS OF PRIVATE LAW 79, 83–91 (Andrew Robertson
and Tang Hang Wu eds., 2009); Andrew Robertson, Justice, Community Welfare and the Duty of Care,
127 LAW Q. REV. 370 (2011); Andrew Robertson, Rights, Pluralism and the Duty of Care,inRIGHTS AND
PRIVATE LAW 435 (Donal Nolan and Andrew Robertson eds., 2012).
47 See, e.g.,J
ANE STAPLETON,Duty of Care Factors: A Selection from the Judicial Menus,inTHE LAW OF
OBLIGATIONS:ESSAYS IN CELEBRATION OF JOHN FLEMING 59 (Peter Cane and Jane Stapleton eds., 1998); Jane
Stapleton, The Golden Thread at the Heart of Tort Law: Protection of the Vulnerable,24A
UST.BAR REV.
135, 135 (2003) (“Oddly, some Australian lawyers still believe that there is a meaningful tension
between principle and policy as bases for legal reasoning in tort cases. But I have yet to hear a
compelling account of the difference between principle and policy”). See also Jenny Steele, Scepticism
and the Law of Negligence, 52 C.L.J. 437, 441 (1993) (doubting the stability and defensibility of the
principle/policy distinction).
48 See,e.g.,R
OBERT STEVENS,TORTS AND RIGHTS (2007) ch 14; ALLAN BEEVER,REDISCOVERING THE LAWOF NEGLIGENCE
(2007).
136 J. Goudkamp and D. Nolan

6 Acts and Omissions
The relatively modest role that Lord Reed carved out for policy factors meant more
space for principle to be deployed. In Robinson, it was the distinction between acts
and omissions that was of paramount importance. As regards this issue, Lord Reed
said:
The distinction between careless acts causing personal injury, for which the law generally
imposes liability, and careless omissions to prevent acts (by other agencies) causing personal
injury, for which the common law generally imposes no liability, is not a mere alternative to
policy-based reasoning, but is inherent in the nature of the tort of negligence. For the same
reason, although the distinction, like any other distinction, can be difficult to draw in borderline
cases, it is of fundamental importance. The central point is that the law of negligence generally
imposes duties not to cause harm to other people or their property: it does not generally impose
duties to provide them with benefits (including the prevention of harm caused by other
agencies).
49
As with other parts of his judgment, this focus on the act/omission distinction was
consistent with Michael, albeit that there the focus on the distinction had militated
against liability, whereas in Robinson it worked in its favour. Lord Hughes reacted
strongly against Lord Reed’s analysis on this point. In particular, while Lord Hughes
acknowledged the significance of the act/omission distinction, he considered that it
could not explain the absence of a duty of care in all of the police liability cases for
two reasons: first, because there were exceptions to the omissions doctrine; and,
second, because “there is no firm line capable of determination between a case of
omission and of commission”.
50
With respect, however, these are not convincing
objections. The existence of exceptions to a no-liability rule does not necessarily (or
even presumptively) rob the rule of force in cases which do not fall within the
exceptions. For example, the outcome of particular cases can plausibly be explained
by reference to the general rule against recovery of pure economic loss in negligence
even though the rule is not absolute. And the idea that there is not a workable
distinction between positive conduct cases and omissions cases, and that “the great
majority of cases can be analysed in terms of either”,
51
is implausible. On the con-
trary, once it is understood that the distinction in play here is between making things
worse for the claimant and not making them better, the distinction works perfectly
well in most cases. This is because, applying this distinction, it is usually enough for
the court simply to ask itself whether removing the defendant from the picture
entirely would have left the claimant in a better position (in which case the defendant
49 [2015] UKSC 2, [2015] AC 1732, [69].
50 Id., [117].
51 Id.
Robinson v. Chief Constable of West Yorkshire 137

made things worse) or in the same position (in which case the defendant did not
make things worse, and hence is being sued for failing to confer a benefit).
52
7 General Duty Tests
In Caparo, Lord Bridge perceptively observed that “there has for long been a tension
between two different approaches”to the duty concept.
53
Under what he called the
traditional approach, “the law finds the existence of the duty in different specific
situations each exhibiting its own particular characteristics”, whereas under “the
more modern approach”, a single general principle is sought “which may be applied
in all circumstances to determine the existence of a duty of care”.
54
Unsurprisingly
Lord Bridge traced the “modern approach”to the speech of Lord Atkin in Donoghue
to which we referred at the start of this article. Indeed, the enunciation of Lord
Atkin’s famous “neighbour principle”can be (and has been) interpreted as a general
test (although it is unlikely that Lord Atkin intended it to be so interpreted).
The Anns test represented the apogee of the modern approach until it was
abandoned in Caparo. However, even in Caparo the tension between the two ap-
proaches was evident. On the one hand, Lord Oliver stressed that the search for a
general duty test was pointless and detrimental –a pursuit of a “will-o’-the wisp”,
which served “not to clarify the law but merely to bedevil its development in a way
which corresponds with practicality and common sense”.
55
Similarly, Lord Bridge
emphasised that the answer to the duty question lay in what he called “the more
traditional categorisation of distinct and recognisable situations”.
56
And yet at the
same time Lord Bridge could not resist the temptation to identify three “necessary
ingredients”for a duty of care to arise, namely foreseeability, “proximity”and that it
be “fair, just and reasonable”for a duty of care to be imposed.
57
Although Lord Bridge
specifically denied that the latter two concepts were “susceptible of any such precise
52 See further John C.P. Goldberg & Benjamin C. Zipursky, The Place of Duty in Negligence Law,64
VAND.L.REV. 657, 695–696 (2001); Donal Nolan, The Liability of Public Authorities for Failing to Confer
Benefits, 127 LAW Q. REV. 260 (2011); Sandy Steel, Rationalising Omissions Liability in Negligence, 135
LAW Q. REV.484 (2019). Note that in some positive conduct cases the defendant’s negligence will not
leave the claimant worse offbecause the conduct was not a but-for cause of the claimant’s damage,
but merely contributed to its occurrence. See further Jane Stapleton, Unnecessary and Insufficient
Factual Causes,J.T
ORT LAW (forthcoming).
53 [1990] 2 AC 605 (HL) 616.
54 Id.
55 Id., 633.
56 [1990] 2 AC 605 (HL) 618. See also at 628 (Lord Roskill).
57 Id., 617–618.
138 J. Goudkamp and D. Nolan

definition as would be necessary to give them utility as practical tests”,
58
these three
ingredients led in time to the resurrection of the modern approach, under the guise
of the so-called “three-stage test”named (ironically) after the very case which had so
forcefully sought to restore the traditional approach. Robinson and Michael repre-
sent a concerted effort by the court of final appeal to stamp out the general test
approach once and for all. They involve, consequently, a dramatic revival of tradi-
tional duty methodology.
Which approach is preferable? The notion that it is possible to devise a universal
formula that can simply be applied to any case to yield the correct result in relation to
the duty issue certainly has at least superficial appeal, and this presumably explains
why many judges have long sought to come up with one. Further, time-poor judges
likely find general tests to be convenient. Instead of embarking on the potentially
labour-intensive exercise required by the traditional approach (which, particularly
as set out in Robinson, entails an open-ended search for and consideration of anal-
ogies), they can simply input the relevant variables into a formula and have the
answer to the duty problem materialise.
Conversely, there are powerful reasons to resist the allure and “beguiling
simplicity”
59
of general duty tests. One is that in order to encompass a sufficiently
wide range of possible facts, the concepts that constitute a general duty test are
generally so abstract (“neighbourhood”,“proximity”and so on) that in themselves
they provide little or no guidance as to whether a duty of care ought to be recog-
nised.
60
At worst, these concepts are not only empty of substantive content, but they
also conceal the true path along which judges reason.
Another and more profound problem with general duty tests is that they
necessarily encroach on the role of precedent. Indeed, taken literally, general tests
seem to leave no room for the doctrine of precedent at all. The judge simply applies
the test to the facts and the test itself produces the right answer. Earlier authority
simply goes by the board, as in the worst excesses of the Anns era. In practice, of
course, a more nuanced approach is taken, but the basic question of how any
particular general test is to be reconciled with the doctrine of precedent never
receives a satisfactory answer, because no such answer can be given. Hence the use
of a general test in combination with traditional precedent-based reasoning neces-
sarily generates irresolvable tensions and the law inevitably becomes unclear and
incoherent.
58 Id., 618.
59 MAURICE MILLNER,NEGLIGENCE IN MODERN LAW 236 (1967).
60 For a well-known essay critiquing the concept of proximity and which played an important role in
preventing the Caparo test from taking root in Australia, see MICHAEL MCHUGH,Neighbourhood,
Proximity and Reliance,inE
SSAYS ON TORTS (Paul D Finn ed., 1989).
Robinson v. Chief Constable of West Yorkshire 139

Suppose, for example, that we take seriously the proposition put forward by the
Court of Appeal in Robinson that the three-stage Caparo test be used in every duty
case. Where would this leave the complex web of duty rules that the courts have
developed to govern liability for, for example, psychiatric injury?
61
Does the court
simply ignore the relevant decisions, despite the fact that they appear to govern the
claim in question, and apply the Caparo test instead? And if the answer is in the
negative, meaning that in psychiatric injury cases precedents must be followed, what
reason is there for a special approach in that particular enclave?
The disregard of precedent is inherently objectionable in a system that is
founded on the doctrine of stare decisis. However, it is possible to distil two particular
problems that it causes. The first is that it necessarily leads to ad hoc decisions and, in
turn, uncertainty in the law. The outcome of cases becomes less predictable,
62
which
in turn generates more litigation. Indeed, it might be thought that the suggestion of
one senior judge that a general duty test is required if the law is not to become a
“morass of single instances”
63
gets things backwards and that, on the contrary, that is
exactly what general tests are prone to produce. A second shortcoming of general
tests is that a judge who is both freed from the constraints of precedent and asked to
apply a test made up of concepts so abstract as to be meaningless (as the general duty
tests formulated by the English courts have invariably been) is essentially given carte
blanche to decide the case however he or she wishes. Although this does not
necessarily result in bad outcomes in individual cases, such an approach is anti-
thetical to the philosophy of the common law and weakens the authority of the
decisions that are made.
It would be an unconvincing response to the critique that has been offered to say
that a general test such as the Caparo test is only to be used in novel cases. One
problem with this reply is that even in novel cases there will always be relevant
authorities from which guidance can be gleaned by analogy, and as we have said
such authorities cannot properly be accommodated within the rubric of the Caparo
test. In an advanced system of precedent-based rules, there is simply no such thing as
a jurisprudential vacuum, where a court (even at the ultimate appellate level) is
completely unconstrained by earlier case law. Although they may be stronger or
weaker in individual cases, analogies always exist. To this it can be added that there
is, in reality, no bright line between categories of case in which the duty issue has
already been resolved and novel cases because the use of analogical reasoning blurs
the boundary. And since it is only after the authorities have been fully considered
61 See supra note 14.
62 See Jonathan Morgan, The Rise and Fall of the General Duty of Care, 22 P.N 206, 217–218 (2006).
63 Customs and Excise Commissioners v. Barclays Bank plc [2006] UKHL 28, [2007] 1 AC 181, [8] (Lord
Bingham).
140 J. Goudkamp and D. Nolan

that any assessment of the novelty of the claim before the court can be made, the idea
that the Caparo test could be used ab initio in novel cases makes little sense.
By contrast, the traditional approach, revitalised by Robinson, simply involves
standard common law reasoning.
64
That is to say, it entails working with the existing
authorities to determine whether any particular case is covered by binding prece-
dent, and, if not, reasoning by analogy and by reference to relevant considerations in
order to decide whether or not a duty of care is owed.
65
By reducing and constraining
the degree of individual judgment involved in deciding cases, it brings the law of
negligence closer to the “rule of law ideal where the outcome of the case does not
depend upon the individual views of the particular decision-maker, but upon the
common understanding of what the law requires”.
66
While the exercise required by
the traditional approach may be more or less straightforward in a particular case,
and although there might be legitimate disagreement about the considerations a
court should take into account, there is nothing mysterious about what it entails.
It might be said in response to the foregoing that the traditional approach is itself
just another general test. However, although the courts have sometimes presented
the incremental methodology endorsed in Robinson as though it exists on the same
plane as the Caparo test,
67
this is a clear error. The whole point of the traditional
approach is that there is, and can be, no such thing as a general duty test.
68
When the
Supreme Court in Robinson rejected the existence of the Caparo test, it did not install
the incremental approach to fill the void that it created. Rather, no test for the
existence of a duty of care was adopted.
It is sometimes suggested that the traditional approach is stultifying and does not
allow the tort of negligence to develop.
69
However, this is a caricature apparently
borne of the failure to appreciate that the traditional approach is just standard
common law reasoning. After all, no-one could seriously suppose that the common
law itself is incapable of development just because it is a system of rules based on
precedent and reasoning by analogy. Conversely, it is true that applying the tradi-
tional approach the courts are less likely to make radical changes to the
64 See KEITH STANTON,Incremental Approaches to the Duty of Care,inTORTS IN THE NINETIES 34, 39
(Nicholas J. Mullany ed., 1997) (“incrementalism may be regarded as restoring a traditional method of
adjudication to the tort of negligence”).
65 See also Plunkett, supra note 44, 71.
66 Grant Lamond, Analogical Reasoning in the Common Law, 34 O.J.L.S. 567, 576 (2014).
67 See, e.g.,Customs and Excise Commissioners v. Barclays Bank plc [2006] UKHL 28, [2007] 1 AC 181,
[7].
68 See Crimmins v Stevedoring Industry Finance Committee [1999] HCA 59, (1999) 200 CLR 1, [73]
(McHugh J stating that “[t]he policy of developing novel cases incrementally by reference to analo-
gous cases acknowledges that there is no general test for determining whether a duty of care exists”).
69 See, e.g., David Howarth, Negligence after Murphy: Time to Re-think, 50 C.L.J. 58, 71 (1991).
Robinson v. Chief Constable of West Yorkshire 141

circumstances in which a duty of care is or is not owed, but this is simply an incident
of the common law process itself.
8 Conclusion
The decision in Robinson has the potential to usher in a new era in the troubled
history of the duty of care element of the tort of negligence, an era marked by greater
predictability and consistency. The decision elevates the importance of precedent
and principle in the determination of whether a duty of care is owed, and diminishes
the significance of policy (while still according it a subsidiary role in cases which are
not governed by established authority). Relatedly, it rejects as misconceived the idea
that all duty cases can be resolved by the use of a general duty test. Instead, the
methodology of the common law itself provides the answer to novel duty cases.
142 J. Goudkamp and D. Nolan