The Negligence Liability of Public Authorities For Omissions
The Negligence Liability of Public Authorities For Omissions
545–569
doi:10.1017/S0008197319000692
TOM CORNFORD*
I. INTRODUCTION
A. The Problem: The Relationship Between the Omissions Principle and
Public Authority Liability
Suppose that a public authority fails to protect a citizen against some harm
from which it is the function of the authority to protect her. Suppose, for
example, that a police force fails to protect a citizen from a crime or that
the fire brigade fails to come to the aid of a citizen whose house is burning.
Does the fact of the authority’s failure being an omission exempt the
authority from liability in negligence in just the way it would if the failure
was that of a neighbour who had the capacity to render assistance but failed
to do so? In other words, is the authority exempt from liability by reason of
the omissions principle, the principle, that is, that the common law does not
impose liability for omissions? If, conversely, a court were to hold the
authority liable, would this represent the abrogation of the principle?
Would it mean that a public authority could be held liable wherever it
* Address for Correspondence: School of Law, University of Essex, Wivenhoe Park, Colchester, CO4
3SQ, UK. Email: tomc@essex.ac.uk.
545
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
546 The Cambridge Law Journal [2019]
had the capacity to prevent some harm to a citizen and failed to do so? Or
would it represent some special exception to the principle?
The answer I shall give in this article is that whether or not the omissions
principle applies to public authorities as it does to private persons depends
upon the answer given to a prior question: can a duty of care be based upon
the public law powers and duties of a public authority? If it can, then where
the claim made is that the authority has failed in the performance of one of
its assigned functions, the rationale that supports the omissions principle
ceases to apply and the authority can be liable for omissions as well as
acts, regardless of whether the omissions complained of fall within any
of the principle’s established exceptions. If it cannot, then the rationale
that supports the omissions principle applies just as it would to a private
person and only where the omissions complained of fall within one of
the established exceptions will it be possible for liability to arise. The
aim of the article is thus deflationary or prophylactic. It attempts to remove
the omissions principle from the prominent position given to it by some
commentators and to redirect attention to what I contend is a more import-
ant issue: the relationship between the public law powers and duties of pub-
lic authorities and the duty of care.
1
Michael v Chief Constable of South Wales [2015] UKSC 2, [2015] 2 A.C. 1732.
2
See especially [97]–[102].
3
S. Tofaris and S. Steele “Negligence Liability for Omissions and the Police” [2016] C.L.J. 128. Like a
number of other writers on the subject, Tofaris and Steel define the omissions principle so as to include a
list of standard exceptions. Their definition, which was approved by Lord Kerr and Lady Hale in Michael
is as follows: “[i]n the tort of negligence, a person A is not under a duty to take care to prevent harm
occurring to person B through a source of danger not created by A unless (i) A has assumed responsi-
bility to protect B from that danger, (ii) A has done something that prevents another from protecting B
from that danger, (iii) A has a special level of control over that source of danger, or (iv) A’s status created
an obligation to protect B from that danger.” In this article, because I wish to emphasise the detachability
of the exceptions from the basic prohibition, I use the expression “omissions principle” to refer only to
the latter.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
C.L.J. The Negligence Liability of Public Authorities For Omissions 547
Stovin v Wise4 called the “why pick on me?” argument, namely the argu-
ment that where there is a large or indeterminate class of people able to ren-
der assistance to a person in need, there is no reason why one member of
the class rather than another should be held liable. Neither of these justifi-
cations works when applied to public authorities. Public authorities are not
entitled to personal autonomy and where they exist for the very purpose of
rendering the kind of assistance that a person in need requires, there is a
good reason for picking on them.
Yet opposed to this is the view advanced by Nolan in an article entitled
“The Liability of Public Authorities for Failing to Confer Benefits”.5 Nolan
argues that the distinction between acts and omissions is a fundamental one
that applies irrespective of whether the defendant is a private person or a
public authority and that the tendency to neglect the issue and to concen-
trate instead on the role of public law principles in negligence or on the val-
idity of policy arguments is a mistake.6 Nolan is not alone in asserting that
the rule against liability for omissions is, by itself, a fundamental barrier to
attaching liability to a public authority in a case in which it causes harm by
an omission. In her article “In Defence of the Omissions Rule in Public
Authority Negligence Claims”,7 Hannah Wilberg argues that the rule
(or principle as I have called it) applies in the same way to private and pub-
lic defendants. Elsewhere, one finds at various places in the literature the
assertion that to allow liability for omissions in cases involving public
authorities will undermine the omissions principle and lead to a state of
affairs in which each of us owes a duty “to all the world”.8
The view I express in this article has implications both for the arguments
advanced on this issue by Nolan and for those of Tofaris and Steel. It
means, on the one hand, that the omissions principle does not have the pri-
ority ascribed to it by Nolan. A claim based on harm caused by a public
4
Stovin v Wise [1996] AC 923 (HL).
5
D. Nolan, “The Liability of Public Authorities for Failing to Confer Benefits” (2011) 127 L.Q.R. 260.
6
In the most recent edition of their tort textbook, Nolan and Oliphant make references to the Tofaris and
Steel and to the Nolan article referred to above and ask roughly the question of the present article:
“should the omissions principle apply with the same force in cases involving public authority defendants
as in cases involving private defendants?” See M. Lunney, D. Nolan and K. Oliphant Tort Law: Text and
Materials, 6th ed. (Oxford 2017), 546–47.
7
H. Wilberg, “In Defence of the Omissions Rule in Public Authority Negligence Claims” (2011) 19 T.L.J.
159.
8
See e.g. J. Smith and P. Burns “Donoghue v Stevenson – The Not So Golden Anniversary” (1983) 46 M.
L.R. 147, at 156; J. Sopinka, “The Liability of Public Authorities: Drawing the Line” (1993) 1 Tort Law
Rev. 123, at 148–49. One set of views that I shall not consider in this article are those of rights theorists
such as Robert Stevens and Allan Beever. These authors argue that, beyond the commonly accepted list
of exceptions to the omission principle, there is no liability for omissions because we do not have rights
that others act so as to assist us. To rebut their arguments would require more prolonged treatment than I
can give them here but in a nutshell, they are quite unable to supply arguments to support their view of
which rights we have. For Stevens and Beever’s views on omissions, see R. Stevens, Torts and Rights
(Oxford 2007), 9; A. Beever, Rediscovering the Law of Negligence (Hart, 2007), ch. 9. For critique of the
rights theorists attempts to justify their view of the rights we have, see T. Cornford, “Public Authority
Liability and the Heteronomy of Tort Law” (2013) 21 T.L.J. 16, at 22–30.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
548 The Cambridge Law Journal [2019]
authority’s failure to act cannot be rejected purely on the basis that the fail-
ure is an omission falling outside the established exceptions. The courts
must first be prepared to affirm – as they have done since the Gorringe
case9 – that a duty of care cannot be based on a public authority’s public
law powers and duties or, if they are not prepared to affirm this, then that
a duty of care based on the authority’s public law powers and duties is
excluded on policy grounds. So far as the arguments of Tofaris and Steel
are concerned, my view entails that their promotion of the view that the
omissions principle does not apply to public authorities is only supportable
if it is seen as based on a more far-reaching proposition: that a duty of care
can be based on a public authority’s public law powers and duties. This
entails in turn, that the restrictive position taken in Gorringe is mistaken
and that the courts should return to allowing the possibility of claims
based on public authorities’ public law powers and duties as they did for
the quarter-century before Gorringe was decided. As we shall see,
Tofaris and Steel stop some way short of this conclusion.
9
Gorringe v Calderdale Borough Council [2004] UKHL 15, [2004] 1 W.L.R. 1057.
10
T. Honoré, “Are Omissions Less Culpable?” in P. Cane and J. Stapleton (eds.), Essays for Patrick Atiyah
(Oxford 1991), 31ff.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
C.L.J. The Negligence Liability of Public Authorities For Omissions 549
just to the extent that it is possible to impose on them duties of care con-
tinuous with their public law powers and duties.11
In Gorringe, Lord Hoffmann, with whom the other members of the House
of Lords agreed said: “I find it difficult to imagine a case in which a com-
mon law duty can be founded simply upon the failure (however irrational)
to provide some benefit which a public authority has power (or a public law
duty) to provide.”12 Lord Scott went further:
In my opinion, if a statutory duty does not give rise to a private right to sue for
breach, the duty cannot create a duty of care that would not have been owed at
common law if the statute were not there. If the policy of the statute is not con-
sistent with the creation of a statutory liability to pay compensation for damage
caused by a breach of the statutory duty, the same policy would, in my opin-
ion, exclude the use of the statutory duty in order to create a common law duty
of care that would be broken by a failure to perform the statutory duty. I would
respectfully accept Lord Browne-Wilkinson’s comment in X (Minors) v
Bedfordshire County Council . . . that “the question whether there is such a
common law duty and if so, its ambit, must be profoundly influenced by
the statutory framework with which the acts complained of were done.” But
that comment cannot be applied to a case where the defendant had done noth-
ing at all to create the duty of care and all that is relied on to create it is the
existence of the statutory duty. In short, I do not accept that a common law
duty of care can grow parasitically out of a statutory duty not intended to
be owed to individuals.13
To see how a common law duty of care can be founded on the failure to
provide some benefit which a public authority has a power or duty to pro-
vide – or “grow parasitically” out of such a power or duty – it is helpful to
look back to the case to which Gorringe was a kind of sequel, Stovin v
Wise.14 The latter, it will be recalled, concerned a road accident at a junc-
tion. The plaintiff suffered injury when a car emerging from a side road
knocked him from his motorcycle. Part of the reason for the accident
was that at the junction in question, the motorist’s vision was obscured
by the presence of a mound of earth. The plaintiff brought his action against
both the motorist and the highway authority, Norfolk County Council. The
latter had powers under the Highways Act 1980 to serve a notice requiring
11
Just to be clear, I shall not argue that courts should impose on public authorities duties continuous with
their public law powers and duties. I do in fact think that they should but in this article, for reasons of
space, I confine myself to the argument set out in the text above.
12
Gorringe [2004] UKHL 15, [2004] 1 W.L.R. 1057, at [32].
13
Ibid., at para. [71].
14
Stovin [1996] AC 923 (HL). In this sentence I talk simply of a public authority’s powers and duties. In
the two passages quoted, Lord Hoffmann talks of “public law” duties and Lord Scott of “statutory”
duties. I address the question of the relationship between public and statutory powers and duties below.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
550 The Cambridge Law Journal [2019]
the occupier of the land on which the mound was situated to remove it in
order to improve the safety of the junction. It had, in fact, planned to do this
work itself but had failed, apparently through pure inadvertence, to execute
its plan.
The case raised in an especially clear form the problem of whether a duty
of care can be founded on a public authority’s public law powers and
duties. If the highway authority were treated as if it were a private person
then there were no grounds for imposing a duty of care. The Highways Act
provided the authority with a panoply of powers for dealing with dangers,
obstructions and inconveniences on the highway and the power to order
removal of the mound was one of these. It also had a duty under the Act
to maintain the highway itself. But there was no explicitly stated duty to
remove obstructions to vision at junctions nor any more general explicit
duty to deal with dangers and obstructions. On the facts of the case,
there was thus no duty of the sort that could form the basis of an action
for breach of statutory duty.
In circumstances of this type, the only justification for imposing a duty of
care is that the common law duty of care gives effect to a public law duty,
the latter not to be found in terms in the governing statute but itself imposed
upon or ascribed to the authority on the basis of an exercise in statutory
interpretation undertaken by the court. In a case involving a highway
authority’s power to remove dangers, obstructions and inconveniences on
the highway, such as Stovin, the exercise would involve something like
the following. A highway authority has many such powers. Clearly it is
given these powers for a purpose. It is not at liberty to shirk their use
and to simply ignore problems on the highway that they might be used
to solve. One way of expressing this would be to say that the authority
has an overarching duty to make the highway safe, even though no such
duty is expressed in terms in the governing statute.15 A duty of this sort
would be a target duty leaving a great deal of discretion to the authority
as to how to fulfil it. Nonetheless, in the light of this overarching duty or
purpose, a court will sometimes feel it appropriate to find that the authority
ought to exercise one of its powers, generally on the ground that not to do
so would be irrational. On the facts of Stovin itself, it was arguably
irrational for the authority, having identified the danger at the junction
and having put on foot a scheme for its removal, not to do so. On this
basis, it would have been possible to find a public law duty to complete
the scheme but also, since not completing the scheme entailed the
15
In postulating the existence of such a duty one would also take into account powers and duties in other
relevant statutes such as the duty on any highway authority that is a local authority in s. 39 of the Road
Traffic Act 1988 to carry out studies into road accidents in its area and to take such measures as appear to
it to be appropriate to prevent them.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
C.L.J. The Negligence Liability of Public Authorities For Omissions 551
16
Stovin [1996] AC 923 (HL), 953 D–E.
17
Ibid., at p. 936.
18
Ibid., at p. 936 B.
19
Ibid., at p. 937.
20
Ibid., at pp. 939–41.
21
Kent v Griffiths [2001] Q.B. 36.
22
Larner v Solihull Metropolitan Borough Council [2001] R.T.R .32.
23
Gorringe v Calderdale MBC [2002] EWCA Civ 595, [2002] R.T.R. 27.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
552 The Cambridge Law Journal [2019]
24
Anns v Merton Borough Council [1978] A.C. 728.
25
Stovin [1996] AC 923 (HL), 931 F.
26
Ibid., at p. 933 F–G. Cf. Lord Bingham in his essay “The Uses of Tort Law” (2010) 1 J.E.T.L. 3: “if a
member of the public whom a public service exists to serve suffers significant injury or loss through the
culpable fault or reprehensible failure of that service to act as it should, is it not consistent with ethical
and, perhaps, democratic principle that the many, responsible for funding the service, should bear the
cost of compensating the victim?”
27
As they did most notably in Hill v Chief Constable of West Yorkshire [1989] A.C. 53 and X (Minors) v
Bedforshire County Council [1995] A.C. 633.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
C.L.J. The Negligence Liability of Public Authorities For Omissions 553
this, it would go something like the following. Public authorities are given
special powers in order to serve the public, both the general public as a
whole and, in some cases, individual members of it. A special body of
law, public law, exists both to facilitate the performance by public author-
ities of their functions and to hold them to account by ensuring both that
they fulfil their duties and that they do not exceed their powers.
Where a public authority harms an individual citizen either by acting in
excess of its powers or by failing to fulfil a duty that it owes to an individual
citizen, the same remedy is appropriate as would be appropriate if one pri-
vate person had harmed another, namely compensation or damages. In a
system which has not traditionally made a clear distinction between public
and private law and in which private law proceedings can be used to ques-
tion the public law lawfulness of acts and omissions of public authorities,
there is no reason not to use the law of tort to provide this remedy.
To defend this justification and to draw out all the implications of the
approach it supports would require more space than is available here and
since my purpose in outlining it is to describe its relation and that of its
opposite to the omissions principle, I turn next to the omissions principle’s
rationale.28 Before I do so, however, two features of the approach are worth
touching on since they are relevant to the argument that follows. First, I
mentioned above the courts’ practice in the years following Anns of elim-
inating by means of policy considerations the possibility of a duty of care
that Anns seemed to promise. The public law-based approach considered
here is not, however, incompatible with the use of policy considerations
to limit the incidence of the duty of care as long as this is genuinely
done to prevent the bad consequences that imposing a duty of care might
have in particular cases rather than to thwart the approach altogether.
Second, since the justification adverts to the special powers possessed by
public law bodies generally there is no need to think of the approach as
being exclusively concerned with public bodies’ statutory powers and
duties: where, as in the case of the Crown or the police, a body had sign-
ificant common law public powers, the approach could licence the deriv-
ation from these of a duty of care.
28
I have set out in much greater detail an approach and justification like that describe in the text in my
book, Towards a Public Law of Tort (Aldershot 2008).
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
554 The Cambridge Law Journal [2019]
something that could not give rise to liability.29 In order, therefore, to give
the rule substance it must be possible to identify some neutral distinction
between action and inaction that is independent of the rationale.
The key non-normative notion is that of inaction, in the sense of absence
of bodily movement, that gives rise to harm, in the sense that it is possible
to say that a given instance of harm would not have occurred if the person
whose inaction is in issue had acted. This notion encompasses both the case
of the passer-by who does nothing to save a child drowning in shallow
water and of the person who does not take advantage of the opportunity
to walk up to a stranger in the street and offer him a $100 bill.30 An omis-
sion is a case of inaction in relation to which there is some reason to assert
that the non-actor ought to have acted. In relation to the examples just
given, the first but not generally the second would be thought of as an omis-
sion since most of us would recognise, at least, a moral duty to assist a
dying child while most of us would deny the existence of an obligation
to give money to random strangers.31
29
Honoré, “Are Omissions Less Culpable?”, p. 37.
30
To borrow an example from Joel Feinberg, The Moral Limits of the Criminal Law: Volume 1, Harm to
Others (Oxford 1984), 135.
31
In seeking a basic non-normative concept as a building block for discussion of the omissions principle, I
eschew the distinction between harming and failing to confer a benefit. Some writers who insist on the
absoluteness of the omissions principle promote this distinction as a straightforward factual one use of
which makes it possible to avoid the difficulties associated with the act/omission distinction: see for
example Nolan, “The Liability of Public Authorities”, p. 260; M. Bowman and S. Bailey, “Negligence
in the Realms of Public Law – a Positive Obligation to Rescue?” [1984] P.L. 277, at 283. On this
view, in order to determine whether a person’s blameworthy behaviour is an act or an omission, we
need only ask whether she has worsened or merely failed to improve the victim’s position. The objection
to this view is that we cannot divorce the question of whether someone is improving or worsening
another’s situation from the question of what that other person is entitled to expect. If A gives B some-
thing she is not entitled to expect, we may say that A is conferring a benefit on B or improving B’s situ-
ation. If A gives B something that she is entitled to expect, it would not be natural to say A is improving
B’s situation. A will be simply maintaining B’s current condition. Everything turns on what B is entitled
to expect and determining this involves all the difficulties that are involved in determining whether B’s
inaction amounts to an omission. The harm/failure to confer a benefit distinction is often used, moreover,
in a way that is misleading. On the one hand, to speak of an intervention to assist someone as “conferring
a benefit” tends to suggest that the intervention is something to which the recipient of the assistance is not
entitled. On the other hand, users of the notion are in the habit of describing any omission not falling
within one of the traditional exceptions to the omissions principle as a failure to confer a benefit.
Bailey and Bowman characterise the defendant public authority’s failure in Anns v Merton Borough
Council in this way, for example, and Nolan does the same in relation to Gorringe. Users of the harm/
failure to confer a benefit distinction thus skew discussion in favour of the view that omissions not falling
within one of the traditional categories of exception to the omissions rule cannot give rise to liability while
at the same time pretending that the distinction has a clear factual basis when it does not.
32
Stovin [1996] AC 923 (HL).
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
C.L.J. The Negligence Liability of Public Authorities For Omissions 555
the second is what Lord Hoffmann calls the “why pick on me” argument,
namely the fact that there often appears to be no principled way of confi-
ning the class of persons who should be subject to an affirmative duty or
of deciding which members of the class should bear it.33
Honoré attempts to explain the omissions principle in utilitarian terms.34
He suggests that we sanction acts more severely than omissions because
abstaining from harmful action requires less effort than acting positively
to assist others while the effects of abstention are more widely distributed
than the effects of positively providing assistance. Thus it is easy to refrain
from poisoning others and it benefits everyone, whereas for an individual to
provide food to a starving person might be onerous and would benefit only
the individual assisted. The implication is that the net benefit is greater
when we require people to refrain from harmful action than it is when
we require people to assist others. It is not hard, however, to think of situa-
tions where this is not the case. A rule requiring ships in fog to sound their
horns presumably has greater net benefit than a rule that simply requires
them to avoid hitting each other. Kortmann invites us to consider the
example of a busy road blocked by an unconscious person.35 The total
social cost of requiring one driver to remove the unconscious person is
lower than the social cost of requiring each driver to steer round her.
What does seem clear is that subjecting each of us to myriad duties to
help others would create an unsustainable burden. This brings us back to
Lord Hoffmann’s explanations of the principle. The second of these is use-
fully elaborated by Feinberg who points out that requiring each of us to ren-
der assistance whenever we were able to would not only impose an
33
His Lordship added a third, economic justification. This was that those carrying out harm-causing activ-
ities were required to internalise the costs they imposed on others and that finding liability was a way of
making them do so. This made sense, his lordship said, where a defendant positively imposed a cost by
harmful action but not where, by failing to act, she failed to confer a benefit. It is doubtful whether
English tort law can be explained in economic terms (although see R. Posner and W. Landes, The
Economic Structure of Tort Law (Cambridge, MA 1987)). But putting this to one side, the dominant
school of economic analysis of law does not insist that each activity internalise the costs that, on the
basis of traditional legal notions, we would characterise a defendant as imposing on others. Instead, it
requires that the necessary precautions or change in behaviour be undertaken by whomever can do so
most cheaply. So, for example, if fumes from a factory were damaging the health of neighbouring house-
holders and this damage could be avoided most cheaply by the householders staying indoors and keeping
their windows closed, the economic approach would forbid a finding that the factory was liable. If the
person who should take steps to avoid accidents or other social harms is the “cheapest cost avoider”, it is
hard to see why this person should not be someone who could positively intervene so as to prevent the
harm as well as someone who could avoid the harm by refraining from action. Of course, to require each
person to assist another where her doing so was cheaper than requiring the creator of the harm to take
greater care or the victim to bear the loss might create an unmanageable burden and so there would have
to be rules to decide under what circumstances a given person should be under an affirmative duty. This
just takes us back to the Feinberg argument described in the text above. For basic exposition of the eco-
nomic approach see e.g. R.A. Posner, Economic Analysis of Law, 9th ed. (New York 2014), chs. 1, 6;
G. Calabresi, The Cost of Accidents (New Haven 1970); P. Cane, Atiyah’s Accidents, Compensation and
the Law, 8th ed. (Cambridge 2013), 435–55. For a discussion of the application of the economic
approach to the problem of omissions, see J. Kortmann, Altruism in Private Law: Liability for
Nonfeasance and Negotiorum Gestio (Oxford 2004), 18–23.
34
Honoré, “Are Omissions Less Culpable?”, pp. 31–32.
35
Kortmann, Altruism in Private Law, pp. 19–20.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
556 The Cambridge Law Journal [2019]
36
Feinberg, The Moral Limits, pp. 169–71.
37
Honoré, “Are Omissions Less Culpable?”, p. 33.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
C.L.J. The Negligence Liability of Public Authorities For Omissions 557
38
Honoré’s reference to “past dealings” might be taken to indicate that distinct duties are only those that
arise as a result of some form of voluntary undertaking but it is clear from the rest of his essay that he
means the expression to refer to the range of generally recognised exceptions to the omissions principle.
39
Honoré, “Are Omissions Less Culpable?”, p. 33.
40
Ibid., at pp. 48–49.
41
Ibid., at p. 33.
42
As for example in Peter Singer’s claim that we should all help people in poorer countries up to the point
at which the loss to us outweighs the gain to them: see P. Singer, “Famine, Affluence and Morality”
(1972) 1 Phil.& Pub.Aff. 229.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
558 The Cambridge Law Journal [2019]
43
It is not my intention, however, to argue that there is a case for a duty of easy rescue. The argument I
wish to make in this article does not depend on showing that there is. My argument is, rather, that in
order to support a duty of easy rescue it is not necessary to undermine the omissions principle but merely
to make an exception to it; and that, in the say same way, the best argument for public authority liability
for omissions falling outside the established exceptions involves making another exception to the prin-
ciple rather than undermining it.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
C.L.J. The Negligence Liability of Public Authorities For Omissions 559
44
Barrett v Enfield LBC [2001] 2 A.C. 550 (HL) and Phelps v Hillingdon LBC [2001] 2 A.C. 619 (HL) are
examples of the kind of case in which the defendant authority can in principle be held liable for an omis-
sion since the omission in question falls within one of the standard exceptions; and in which, at the same
time, the duty of care that the authority would be thereby held to have breached is entirely consistent with
the authority’s public law obligations.
45
Anns [1978] A.C. 728.
46
See note 27 above.
47
Contrary to the criticism made by Bailey and Bowman of the reasoning in the Anns case in “Negligence
in the Realms of Public Law”.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
560 The Cambridge Law Journal [2019]
V. COUNTERARGUMENTS
This is not all there is to be said, however, on the question of whether the
omissions principle applies in the same way to private and to public defen-
dants. Neither Nolan nor Wilberg accept that the kinds of reasons I advanced
as justifying the omissions principle – the first and second reasons set out by
Lord Hoffmann in Stovin – provide sufficient explanation. Both insist that the
principle must have some deeper significance that explains why it should cut
across the distinction between public and private. In support of this view,
Wilberg invokes the claim that omissions are causally peripheral. For the
same purpose, both Wilberg and Nolan pray in aid Tony Honoré’s claim
that omissions are in general less harmful than acts. Both these claims are
mistaken. In this section, I attempt to show why.
48
The criteria identified by Wilberg include a set of exceptions to the omissions principle roughly equiva-
lent to the standard set and she counts amongst these the possession by the defendant of special powers
of control over the risk giving rise to the harm to the claimant. She considers the argument that the main
justifications advanced for the omissions principle (the “liberty” and “why pick on me” arguments) do
not apply to public authorities and replies that this objection is overcome by the consideration that, in
appropriate cases, public authorities can be made liable for omissions that fall within her set of excep-
tions, especially the control exception. As noted in the text, however, it is not possible to see why on this
view there is sometimes an argument for making a public authority liable when the same argument
would not be made in relation to an analogously situated private person. Why, for example, might we
argue that the police in Michael should have been liable when we might hesitate to argue the same in
relation to a neighbour who would have been in as good a position to assist but did not do so? See further
the arguments in Section VI below.
49
J. Stapleton, “Duty of Care: Peripheral Parties and Alternative Opportunities for Deterrence” (1995) 111
L.Q.R. 301.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
C.L.J. The Negligence Liability of Public Authorities For Omissions 561
50
Wilberg, “In Defence”.
51
Feinberg, The Moral Limits, p. 180; H.L.A. Hart and T. Honoré, Causation in the Law, 2nd ed. (Oxford
1985), 30, note 4 and accompanying text, 38, 140.
52
See e.g. Hart and Honoré, Causation in the Law, pp. 2, 5, 16, 30–31, 37–38, 50, 59, 62, 81, 127–28,
138–41, 236; Feinberg, The Moral Limits, pp. 165–85; J. Stapleton, “Choosing What We Mean by
Causation in the Law” (2008) 73 Mo.L.Rev. 433, at 435–37; J. Stapleton, “Unnecessary Causes”
(2013) 129 L.Q.R. 39, at 40; J. Stapleton, “An ‘Extended But-for’ Test for the Causal Relation in the
Law of Obligations” (2015) 35 O.J.L.S. 697, at 699, 705, 709. Stapleton convincingly refutes
Michael Moore’s claim (made e.g. in “Causation and Responsibility” (1999) 16 Soc.Phil.& Pol’y 1)
that omissions cannot be causes: see “Choosing What We Mean by Causation”, p. 467.
53
Stapleton, “Duty of Care”.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
562 The Cambridge Law Journal [2019]
who can help themselves or third parties. At the same time, it is desirable to
protect deep-pocketed defendants, often public authorities, from the exces-
sive liability that might attach to them if they were required to protect clai-
mants from their own or third parties’ negligence. As a matter of morality,
to make many people potentially liable for the negligence of a third party
would represent a vast inhibition on freedom. The latter consideration is
simply the first of Lord Hoffmann’s two reasons for the omissions principle.
As we have seen, it can be overridden if a duty of care can be based on a
public authority’s public law powers and duties. The former considerations
are examples of the type of policy argument that can limit the incidence of
the duty of care whether or not it can be based on a public authority’s public
law powers and duties. None of these considerations undermine the basic
proposition that there are reasons for making an exception to the omissions
principle in relation to public authorities that do not pertain in relation to
private persons.
54
Nolan, “The Liability of Public Authorities”, p. 285.
55
Ibid., at p. 284.
56
Ibid., at p. 285.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
C.L.J. The Negligence Liability of Public Authorities For Omissions 563
article that supports the claim, however. We saw above that Honoré makes
a sound and illuminating distinction between the background duties that
each of us owes to everyone else and the distinct duties that arise in particu-
lar relationships: we have background duties to refrain from harming others
but we do not necessarily have background duties to act so as to help
others, whereas we may owe distinct duties to help as well as not to
harm and, in this latter context, an omission may be worse than a harmful
act; distinct duties, moreover, generally take priority over background ones.
This is the solid framework of Honoré’s article. Onto it he grafts the argu-
ment about harmful acts being worse than harmful omissions. What is the
relationship between the framework and the argument that Honoré superim-
poses upon it? Are we to understand that harmful acts are worse than harm-
ful omissions at the level of background duties, at the level of distinct duties
or both?
If the argument is intended to show only that harmful acts are worse than
harmful omissions – and more of a threat to security – at the level of back-
ground duties then it tells us nothing that is not already revealed in the ori-
ginal framework. This can be seen if we consider an example that Honoré
seems to think supports his security argument. He says that someone who
drops litter is more culpable than someone who fails to pick up the same
litter even though the harm caused and the likely effort that would be
required to do the right thing would be the same in each case.57 But here
Honoré is comparing two background duties: the duty that a citizen owes
to society as a whole not to drop litter and the duty that a citizen owes
to society as a whole to pick up someone else’s litter. The greater culpabil-
ity of breach of the former duty can be explained by the second, “why pick
on me” justification for the original point about background duties. We can
see this more clearly if we make a comparison of the relative culpability of
someone who drops litter with someone whose job it is to pick up the litter
and who fails to do so, thus breaching a distinct duty. Here it is much less
clear that a harmful act is worse than a harmful omission.
If, on the other hand, Honoré’s security argument is supposed to show
that harmful acts are worse than harmful omissions where those omissions
are breaches of distinct duties, then it simply contradicts what is said else-
where in the article. Honoré says very little to support the idea that this is
the purpose of the argument. He does give in passing an example consistent
with the assumption that it is so: he suggests that a wage cut is worse than
the failure to receive a pay rise.58 But upon examination this example tends
to support the view, advanced earlier in the article by Honoré himself, that
in the context of distinct duties an omission may be just as bad as an act.
Whether or not receiving a wage cut is worse than not receiving a pay
57
Honoré, “Are Omissions Less Culpable?”, pp. 49–50.
58
Ibid., at p. 35.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
564 The Cambridge Law Journal [2019]
rise depends entirely on the context and what one has been led to expect. If
you work in a company that has been doing badly because of the recession
and the necessity of a pay cut has been discussed and the pay cut then turns
out to be only 5% of your income, you may be relieved. If, by contrast, you
were confidently expecting a Christmas bonus worth 10% of your income
and you have already bought expensive Christmas presents on that assump-
tion then you may be gravely disappointed with a bonus worth only 5% of
your salary. Examples of pairwise comparisons of this sort, demonstrating
that in the context of distinct duties an omission may be as damaging to
security as an act, can be multiplied indefinitely.59
The whole idea that acts represent more of a threat to security than
omissions is, moreover, contradicted by what Honoré says earlier in the
article in establishing the framework. Having set out the view that bodily
movement counts as intervention in the world while non-movement does
not he writes:
But superimposed on this picture of the world is a refinement of it, which
treats disruptions of the normal course of events as similar to interventions
which bring about change, though the change is not now a change in the exist-
ing state of affairs but rather in the normal or expected sequence of events. The
regular course of things may take the form not only of recurrent natural events,
but of regular human conduct such as clearing the street when the snow stops
falling. If this regular or expected conduct fails to occur, then failure counts
not, as the first world-view would suggest, as a non-intervention but rather,
on a wider view of what amounts to change, as a sort of intervention in the
world. If, as is often. . .the case, the break in routine violates a norm (the per-
son who does not sweep up the snow may or may not have a duty to do so), it
is a form of not-doing which amounts to an omission and is a potential ground
of responsibility.60
59
One person is injured by the negligent driving of a motorist, another in a train crash caused by the neg-
ligent failure to maintain the track; a haemophiliac loses blood on one occasion because someone care-
lessly cuts her and on another because she cuts herself and the ambulance she calls fail to arrive; one
landowner suffers the burning of his crops because his neighbour carelessly allows fire to spread from
his land, another suffers the burning of his crops because his neighbour flicks a cigarette on to them
while mending the fence between the two properties; and so on.
60
Honoré, “Are Omissions Less Culpable?”, p. 41.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
C.L.J. The Negligence Liability of Public Authorities For Omissions 565
61
Note in this connection, a passage from Honoré’s earlier work with Hart, Causation in the Law, p. 37:
“[t]hough what is treated as normal represents in many ways our practical interests and our attitude to
nature, it would be wrong to identify as the normal and so always as part of the “mere conditions” of
events the course of nature unaffected by human intervention. This is an over-simplification, because
what is taken as normal for the purpose of the distinction between cause and mere conditions is very
often an artefact of human habit, custom or condition. This is so because men have discovered that nature
is not only sometimes harmful if we intervene, but is also sometimes harmful unless we intervene, and
have developed customary techniques, procedures and routines to counteract such harm. These have
become second ‘nature’ and so a second ‘norm’. The effect of drought is regularly counteracted by gov-
ernmental precautions in conserving water or food; disease is neutralised by inoculation; rain by the use
of umbrellas. When such man-made normal conditions are established, deviations from them will be
regarded as exceptional and so rank as the cause of harm. It is obvious that in such cases, what is selected
as the cause from the total set of conditions will often be an omission which coincides with what is rep-
rehensible by established standards of behaviour.” The authors’ primary concern in this passage is to
show that omissions can be causes. But the passage also tends to support the view that, generally speak-
ing, harm-causing omissions that take place in the context of distinct duties are no less serious than harm-
causing acts.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
566 The Cambridge Law Journal [2019]
principle but merely to make another exception to it, an exception that can
be justified, as can the more well-established exceptions, on the ground that
in the circumstances in question there is a distinct duty. Nolan’s view of the
omissions principle involves treating the principle together with the well-
established exceptions as a package, one whose contents are fixed and
immutable.62 Hence he thinks that to step outside the principle and the
established exceptions is to reject the principle altogether. He appears not
to recognise that the principle must have a rationale capable of justifying
the exceptions to it as well as the principle itself and that, if this is so, it
must be possible to make new exceptions if they too can be justified by ref-
erence to the rationale. One can only arrive at this position by reifying the
principle and its established exceptions, in other words by treating them as
sacrosanct and only addressing the reasoning that supports them as an
afterthought.
62
Wilberg conceives of the principle in the same way: see Wilberg, “In Defence”.
63
Tofaris and Steel, “Negligence Liability”, pp. 129, 142, 145, 151.
64
See note 3 above.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
C.L.J. The Negligence Liability of Public Authorities For Omissions 567
these duties as continuous with or growing out of the duties and powers that
define their status.
Tofaris and Steel do not use this mode of argument, however, in trying to
explain why the police in the Michael case should have been subject to a
duty of care. Instead they adopt a curious roundabout procedure. They
argue for the existence of a moral duty to take care to prevent reasonably
foreseeable physical injury to others which, they claim, gives rise in turn
to a legal duty.65 Then, in order to explain why this duty applies only to
public authorities and not to private individuals, they refer back to their dis-
cussion of the omissions principle and argue that while the rationale that
underpins it does not exclude imposition of a duty of care on public author-
ities such as the police, it does exclude its imposition on private persons.
Two features of this argument are worth pointing out. First, its roundabout
character is odd. In considering the imposition on a landowner of a duty to
take care to extinguish a fire on his land, or on a bar owner of a duty to pre-
vent a drunk person from driving home, or on an employer of a duty to pro-
vide a safe system of work, it is not common practice to suggest a duty
extending to all persons and then to try to narrow it down by reference to
the fact that the putative defendant has a particular status while other persons
do not. Usually, exemption from the omissions principle and being subject
to a duty of care are both treated as following from the defendant’s status.
Second, in order to succeed, the argument must begin by ascribing to
citizens generally an extremely extensive duty, one equivalent to the duty
that the police are alleged to be under in Michael. Tofaris and Steel
never say in terms that private individuals are under a prima facie duty
to disarm persons carrying out dangerous attacks on third parties. This
must be what they are implicitly assuming, however, because their method
is to begin with a moral duty that applies to everyone and then to whittle it
down until they arrive at the legally enforceable duty on the police that they
are seeking. Nothing in their argument augments by reference to their spe-
cial powers the duty to which the police are subject. It might be said that the
assumed moral duty only applies in the first place to those able to disarm a
dangerous attacker. But suppose a group of strong men were apprised of an
attack taking place somewhere in their neighbourhood: would we consider
them to be subject to a moral duty to intervene as strong as the duty to
which the police are subject? It is worth noting, moreover, that the exist-
ence of the extremely extensive duty I refer to here is deduced from an
argument in favour of a duty of easy rescue of the sort that might be
owed to a child drowning in shallow water. How the authors get from a
duty of easy rescue to the much more onerous duty to protect someone
from a dangerous criminal is never clear.
65
Tofaris and Steel, “Negligence Liability”, pp. 142–45.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
568 The Cambridge Law Journal [2019]
Shortly after setting out the argument described above for a duty on the
police to aid those in danger, the authors attempt to explain how its scope
would be limited by reference to the police’s “special” – namely public –
legal powers.66 The end result is thus the same as that produced by the
method of argument I suggested above was the correct one. One might
ask, however: was it ever plausible to suggest that private individuals are
under a prima facie duty to disarm persons carrying out dangerous attacks
on third parties? And more generally: why was this circuitous method of
argument adopted?
The reason seems to be a fear of mixing private and public law. That this
is so appears from the reasoning Tofaris and Steel use in order to rebut
another argument against finding a duty of care in the case that concerns
them.67 The argument, deriving from another article by Nolan, is that mak-
ing private persons but not public authorities subject to the omissions prin-
ciple would introduce an “alien public/private distinction” into the common
law.68 The authors response to this is to accept the premise of the argument
(that introduction of the “alien” distinction would be undesirable) but to
insist that the differential treatment of public authorities and private persons
is one licenced by private law. But what justification is there for this fear of
mixing public and private? The two are not clearly differentiated in English
law as they are in some other systems69 and collateral challenge, whereby
the public law legality of some act or omission of a public authority is
tested by means of a private law challenge is a familiar phenomenon.70
VII. CONCLUSION
If one assumes that a public authority’s public law powers and duties can
form the basis of a duty of care, then the problem of omissions vanishes.
The omissions principle does not cease to apply, but there are clear grounds
for making an exception to it where an authority has failed to do the very
thing that it exists to do vis-à-vis the claimant. If one assumes that a public
authority’s public law powers and duties cannot form the basis of a duty of
care, then a public authority can only be subject to liability for its omission
to act where the omission falls within one of the established exceptions: the
authority falls, in effect, to be treated just as if it were a private person for
the purposes of identifying the factors that weigh in favour of a duty of care.
66
Ibid., at p. 145.
67
Ibid., at p. 140.
68
D. Nolan, “Negligence and Human Rights Law: The Case for Separate Development” (2013) 76 M.L.R.
286, at 317.
69
Cf. J. Allison, A Continental Distinction in the Common Law (Oxford 2000); D. Oliver, Common Values
and the Public-Private Divide (London 1999); J. Varuhas, Damages and Human Rights (Oxford 2016),
Part 2.
70
As to which, see generally De Smith’s Judicial Review, 8th ed. (London 2018), 3-123–3-132. See also
R. Clayton and H. Tomlinson, Civil Actions against the Police, 3rd ed. (London 2004), ch. 4.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692
C.L.J. The Negligence Liability of Public Authorities For Omissions 569
71
The question of whether a duty of care can be based directly on a public authority’s public law powers
and duties is of course an aspect of a still larger question: should there be a specialised form of admin-
istrative liability in English law. In principle, if the answer to this question is affirmative, then the change
to the current law might be brought about by creating an avenue of redress entirely distinct from the
existing law of tort. Most proposals for creating some form of administrative liability made since the
Anns case, however, have envisaged that at least part of the task would be fulfilled by developing
the law of negligence: see e.g. JUSTICE-All Souls Review, Administrative Justice: Some Necessary
Reforms (Oxford 1988), ch. 11; Law Commission Consultation Paper No. 187, Administrative
Redress: Public Bodies and the Citizen; Cornford, Towards a Public Law of Tort.
Downloaded from https://www.cambridge.org/core. University of Massachusetts Amherst, on 17 Nov 2019 at 21:22:38, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0008197319000692