I. Introduction
In 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 a responsibility to protect B from that danger, (ii) A has done something which 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 creates an obligation to protect B from that danger.Footnote 1 We refer to this rule and the exceptions to it as the omissions principle.
The manner in which the omissions principle is currently applied to the conduct of public authorities, such as the police, generates results which are, at least prima facie, surprising. So far as the tort of negligence is concerned, a team of officers can permissibly stand by whilst a person is being kicked to death on the street. And, as the Supreme Court has recently affirmed, the police owed no duty to a victim of domestic violence killed by her former partner who had made an emergency call to report the imminent threat to her life.Footnote 2
In this article, we aim to vindicate two main propositions. The first is that the general reasons for applying the omissions principle carry little weight in relation to public authorities in general, and in relation to the police in particular. The second is that there is a convincing case for imposing a duty of care upon the police in respect of negligent failures to prevent reasonably foreseeable personal injury to victims in a proximate relationship with the police and that the scope of that duty can be coherently delimited. This duty encompasses cases which would currently be described as cases of pure omission. We argue that this duty is owed in part because of the police's special status; hence exception (iv) should apply to them.
The article is divided into four main sections. Section II considers, and rejects, the rationale against the imposition of a duty of care on the police for failure to prevent crime on the grounds of the omissions principle. Section III considers other arguments against the imposition of a duty of care on the police and finds most of these deficient. Section IV provides a positive rationale for imposing a duty of care on the police in relation to sufficiently proximate victims of a negligent pure omission. Section V demonstrates that the existing framework of negligence liability, once applied in the light of our arguments, achieves a better balance than the existing law between the competing considerations at work.
II. The Omissions Principle as a Rationale for no Duty of Care
In Michael v Chief Constable of South Wales Police, the majority of the Supreme Court held that the fundamental reason the police are not liable for negligently failing timeously to respond to an emergency call with the result that a person is killed, in circumstances where they did not assume responsibility to do so, is because of the general principle that the common law does not impose liability for pure omissions.Footnote 3 Although that principle has been “worked out for the most part in cases involving private litigants … [it is] equally applicable where D is a public body”.Footnote 4
In this section, we consider various arguments for the common law's reluctance to impose liability for omissions except in special circumstances, and for its reluctance to extend the categories of status which attract a legal duty to take positive action. We argue that, whatever the strength of these arguments in relation to private individuals, they are not convincing as arguments for applying this general position in relation to public authorities.
A. Freedom
In Stovin v Wise, Lord Hoffmann said: “… it is less of an invasion of an individual's freedom for the law to require him to consider the safety of others in his actions than to impose upon him a duty to rescue or protect.”Footnote 5
As a justification for the absence of any affirmative duty of care being owed by public authorities, this statement is problematic. First, whilst it may be that a general duty to take care to protect people from injuries caused by other risk sources is more invasive of freedom than a general duty to take care not to cause injury by one's own action,Footnote 6 it is not obvious that this is true of a limited positive duty, arising only in clearly defined circumstances. Honoré gives the example of dropped litter: “… the trouble involved in disposing of a wrapper neatly in a bin is much the same as the trouble involved in picking a wrapper up.”Footnote 7 His point is that it is not always the case that requiring a positive act is more onerous, and hence invasive of freedom, than the corresponding abstention.
Second, it must be questioned how valuable the freedom of a public authority negligently to fail to take steps to assist an identified individual at serious risk of physical injury is. A private individual's freedom arguably has intrinsic value in so far as her having freedom to do various things contributes to her having an autonomous life. By contrast, the value of the state's freedom is purely instrumental: the state's freedom is valuable only in so far as it contributes to the fulfilment of its proper functions. Furthermore, it has been argued that the moral significance of virtuous acts would be diminished if individuals are not legally free to undertake those acts.Footnote 8 Forced virtue is no virtue at all. Yet it is preposterous to claim that the police's non-negligent response to the serious endangerment of a private individual is a matter of moral virtue. Had the police officers in Michael behaved non-negligently, they would not, ipso facto, have been “virtuous”.
B. Lesser Culpability
Honoré argues that omissions which lead to some effect are generally less culpable than acts which lead to the same effect. The basis of his claim is that acts which lead to harm amount to interventions in the world, whilst omissions are failures to intervene in the world. The significance of this is that an intervention which leads to harm makes things worse as opposed to simply failing to make them better. The former thus “threatens not security so much as the expectation of improvement, which is a different but secondary value”.Footnote 9
This argument, even if correct, does not justify the omissions principle. First, even if the security/expectation of improvement distinction helps to explain why there is a moral difference between acts and omissions, it tells us little about the strength of that distinction. Omissive conduct may in particular cases be so highly culpable as to attract liability.Footnote 10 Second, sometimes there is no moral difference in the culpability of acts and omissions. The mother who deliberately starves her child to death has behaved just as culpably as the mother who deliberately poisons her child. Honoré explains such situations as involving “distinct duties” – in this example, the special duty arising between parent and child.Footnote 11 There may be circumstances where the police can be arguably said to owe a “distinct duty” to prevent some harm even absent an assumption of responsibility to do so.Footnote 12
C. Erosion of Individual Responsibility
Another argument for the omissions principle is that we are primarily responsible for what we do and not for what others do, but if we are held legally responsible for failing to prevent the actions of others, then we blur and potentially erode this moral distinction.
The premise of this argument is correct: intuitively, we are primarily responsible for what we do rather than what others do; our responsibility for our own actions is a special one.Footnote 13 It would indeed erode our sense of special responsibility and authorship over our own lives if there were unlimited moral responsibility for outcomes with which we are prima facie unconnected.Footnote 14 But rethinking the omissions principle in its application to the police hardly implies that the distinction between acts and omissions will be left without any moral significance. First, a limited inroad into the omissions principle does not involve wholesale rejection of the act/omissions distinction. Second, acceptance of a limited legal duty in respect of omissions beyond the existing law need not entail that the person held liable is equally as morally responsible as the primary wrongdoer. If this objection were taken to its logical conclusion, it would rule out any form of accessorial liability in tort law – a liability which is always in respect of the wrong of another, primarily responsible person.
D. Absence of a Right to the Conferral of a Benefit
According to Stevens, “the failure to confer a benefit upon someone else does not, alone, constitute the infringement of a right”.Footnote 15 On this analysis, if A, regardless of her identity or ability to prevent the loss, fails to protect B from some injury threatened by C, then A will have failed to confer a benefit upon B. The normative basis of the claim lies in the “premium placed upon our freedom to choose how we live our lives”.Footnote 16 The argument is thus vulnerable to the same objections mentioned in relation to freedom above.
Nolan also seeks to justify the omissions principle by reference to the view that “we do not have a right good against the whole world that others confer benefits on us”.Footnote 17 Without more, however, this is no justification. The fact that we do not and ought not to have such a right good against the whole world need not imply that we ought not to have such a right good against a limited class of well-positioned potential duty-bearers in limited circumstances. More fundamentally, it is a fallacy to argue from the fact that as a matter of positive law there is no such right to the normative conclusion that there ought to be no such right.
It may be argued that it is self-evident that an individual can never have a moral right, enforceable in law, that another confer a benefit upon her. Yet, given that prominent theorists of rights contend that there can indeed be rights that others deliver some positive assistance, at least in limited situations, this is a difficult claim to accept without further argument.Footnote 18
E. “Why Pick on Me?”
In Stovin v Wise, Lord Hoffmann also referred to the “why pick on me?” argument for the omissions principle.Footnote 19 This argument states that it is unfair to single out one person for failing to take positive steps to protect someone when there are other people who similarly failed to do so.
The shortcomings of this argument have often been pointed out.Footnote 20 First, it does not apply where there is only one person who has negligently failed to provide assistance. Second, in cases where multiple people each breach a duty to give positive assistance, it should not be the case that the more the claimant is the victim of a wrong, the worse off she becomes. To put the point slightly differently, the number of wrongdoers should not dilute the responsibility of any particular wrongdoer. Third, there is special reason to single out a public authority in relation to some injury where the public authority has been tasked by statute or otherwise with taking steps to prevent that injury.
III. Other Considerations Against the Imposition of a Duty of Care
The fundamental objection to the liability of the police for failing to prevent a crime which results in personal injury to the claimant, namely that this would be liability for a pure omission, has little weight. Even so, other objections to such liability have been made in the case law and academic literature. In this section, we argue that almost all of these objections are unconvincing and, in section V, that those that have weight do not preclude the imposition of a duty of care under the orthodox framework for determining the existence of a duty of care.
A. The Hill Policy ArgumentsFootnote 21
In Hill v Chief Constable of West Yorkshire, Lord Keith identified four public policy grounds in support of the non-liability of the police in negligence when investigating or suppressing crime.Footnote 22 None of these is persuasive.
1. No improvement in police standards
Lord Keith's first policy reason was that the recognition of a duty of care is unlikely to improve the performance of police functions because they apply their best endeavours motivated by the general sense of public duty. This has fallen out of favour with the courts.Footnote 23 Recent examples of police institutional misconductFootnote 24 vindicate Lord Steyn's remark in Brooks that “nowadays, a more sceptical approach to the carrying out of all public functions is necessary”.Footnote 25
2. Judicial examination of police strategy
The second policy ground was that negligence claims against the police are likely to raise issues touching deeply on the conduct of a police investigation, including “matters of policy and discretion”, which are unsuitable for determination by the courts. Setting limits on the type of police decisions that are open to judicial examination is legitimate, but, from an analytical perspective, this is better done under “justiciability”.Footnote 26 Moreover, Lord Keith's use of the fact that policy or discretionary issues may arise to conclude that it is never appropriate to impose a duty of care on the police when investigating or suppressing crime is questionable. There may be instances in which no policy or discretionary issues arise, so that it would be wrong to deny a duty of care on that ground.
3. Defensive policing
The third policy factor is that the imposition of negligence liability on the police may lead them to exercise their primary function of investigating and suppressing crime defensively, inhibiting them from taking difficult operational decisions and restricting their freedom to act in the interests of the community.Footnote 27
The argument is not without problems.Footnote 28 Its status in the case law is weakened by the fact that, although accepted in some cases involving public authorities,Footnote 29 it has been rejected in others.Footnote 30 In some of these, it has been suggested that liability would actually enhance the overall standard among public authority employees.Footnote 31 Such an uneven application of the defensiveness argument is not justified on logical grounds. This leaves open the existence of special reasons why the police may be more susceptible to defensive practices than other public services, yet none has been provided on empirical grounds. In the absence of this, the courts proceed merely on intuition, which at best trivialises the complexity of collecting and utilising such evidenceFootnote 32 and at worst gravely misrepresents the true position.
Two glaring examples suffice here. First, judicial statements about defensive policing are inconsistent with the evidence provided by police officers in the course of the Independent Review of the Riot (Damages) Act 1886, where they adamantly rejected the argument on deterrence in connection with the prospect of paying compensation under the Act.Footnote 33 Second, in Van Colle/Smith, Lord Carswell remarked that “police officers may quite properly be slow to engage themselves too closely in … domestic type matters, where they may suspect from experience the existence of a degree of hysteria or exaggeration on the part of either or both persons involved”.Footnote 34 This seems to be at odds with existing evidence, which suggests that there is under-reporting rather than over-reporting of domestic violence,Footnote 35 and that the police response remains “not good enough” with the result “that victims are put at unnecessary risk”.Footnote 36
4. Diversion of police resources
The last policy consideration is that having to defend negligence claims would divert human and other resources of the police from their primary function of investigating and suppressing crime.Footnote 37 This is, to some extent, undermined by the fact that it has been found unpersuasive in other cases involving public authorities.Footnote 38 There also seems to be no reason why it should not in principle apply to claims against the National Health Service, yet the courts do not place any weight on it in that context.Footnote 39 At the same time, it should not be readily assumed that the overall impact on public or police resources is negative.Footnote 40 If the imposition of a duty of care encourages the police to act more carefully,Footnote 41 it may end up saving public resources. Victims of serious crime often rely on medical help from the National Health Service and on welfare support from the state; if, by acting more carefully, the police prevent the crime, those costs will be averted. Therefore, even if police resources are detrimentally affected, public resources may benefit overall. Moreover, it may be that imposition of a duty of care can save police resources. For example, litigation may uncover organisational failings missed by internal inquiriesFootnote 42 and in this way enable the police to adopt a more efficient system in the future. The overall problem is that the likelihood of any of these cannot be ascertained without empirical evidence.Footnote 43
B. Other Arguments for No Duty of Care in Pure Omissions Cases
1. Rule of law
It has been argued that, in light of the fact that private individuals enjoy the benefit of the omissions principle, the rule of law demands that public authorities must also be entitled to the benefit of this principle. This typically relies on a Diceyan conception of the rule of law, which requires that “every man, whatever be his rank or condition, is subject to the ordinary law of the realm and amenable to the jurisdiction of the ordinary tribunals”.Footnote 44
There are several difficulties with this view.Footnote 45 First, although it is often presented by private lawyers as if it were a neutral view defending an uncontroversial principle of equality before the law, in truth it implies a controversial, strongly libertarian view of the responsibilities of the state.Footnote 46 This is because, on this view, the state ought to have no more legal obligations than private individuals. This would challenge many of the modern welfare state's legal obligations to its citizens.Footnote 47
Second, it is doubtful that the rule of law, even on Dicey's conception, is truly engaged here.Footnote 48 The rule of law does not prohibit differential treatment of the state and private individuals if there are normatively important differences between private individuals and the state which justify this.Footnote 49 It is hard to believe that there are no normatively important differences between police officers and private individuals with respect to the suppression of crime.Footnote 50 Such differences obtain in virtue of the special functions entrusted to the police and the dependency which those functions create.Footnote 51 Therefore, whether a private contractor to whom security provision has been outsourced partakes of these normative differences depends upon whether the contractor is entrusted with or exercises these special functions.
2. Unsuitability of a private law analysis
Another argument is that because the normative foundations of tort law lie in corrective justice, tort law provides an inapposite normative framework for analysing the liability of public authorities, since the existence and extent of such liability is principally a question of distributive justice.Footnote 52
Corrective justice is, in essence, the view that, if A wrongfully harms B, A thereby comes under a moral duty to repair that harm.Footnote 53 So how could it be that harms inflicted by public authorities could not come within the scope of corrective justice? It must be either that (i) public authorities cannot, whilst exercising public functions, morally wrong private individuals in the relevant sense or (ii) public authorities cannot harm private individuals.
Proposition (ii) clearly cannot be true. Proposition (i) is also untenable. There is no moral magic about going on duty as a police officer.Footnote 54 For instance, if a police officer strikes a suspect to encourage a confession, this is clearly a wrong relevant to corrective justice. A more plausible version of (i) is that public authorities cannot morally wrong private individuals in the relevant sense in so far as they merely omit to prevent harm occurring to those individuals. This depends upon whether one thinks that a person can ever wrong another by omission. If one accepts the commonsensical viewFootnote 55 that one can morally wrong another person by negligently failing to assist them in certain circumstances, then there is no conflict with corrective justice.Footnote 56
DuBois seems to be committed to the original proposition (i). This is because he thinks the wrongs to which corrective justice responds are wrongs between what he calls “normative equals”.Footnote 57 In his view, persons are normatively equal only if they each have a moral entitlement to set their own ends. Public authorities and private individuals are not normative equals, since the former are not morally entitled to set their own ends, but must serve the public.Footnote 58 Thus, a public authority, at least when pursuing a public function, cannot commit a wrong relevant to corrective justice.Footnote 59 This, however, is implausible. It is peculiar that, when people go “on duty”, they lose the capacity to commit ordinary moral wrongs.
Furthermore, it seems to us that public officials are in any event “normative equals” in DuBois's sense. Public officials, as people, are morally entitled to set their own ends. Yet the role which they have assumed imposes certain special moral obligations on them. Similarly, a fiduciary's role imposes stringent obligations to act in the best interests of the principal, but this does not preclude the tort liability of a fiduciary acting in that capacity. The fact that A is morally obligated to serve the interests of another does not entail that A has no moral entitlement to set her own ends in the relevant sense.
The better view is that the liability of public authorities, like that of private individuals, raises issues of both corrective and distributive justice. It raises issues of corrective justice in so far as it is concerned with the enforcement of moral obligations of repair arising out of interpersonal moral wrongs. It raises issues of distributive justice in so far as the judicial decision to grant a legal right of corrective justice is always a question of distributive justice;Footnote 60 it is a question about the distribution of the legal rights to the enforcement of the moral obligation of corrective justice.
3. Availability of alternative remedies
The courts have in the past refused to impose a duty of care on a public authority where the claimant has an alternative remedy.Footnote 61 Although the argument was disapproved in Barrett Footnote 62 and Phelps,Footnote 63 it has occasionally resurfaced and merits further discussion.
One possibility for a claimant who suffers personal injury as a result of a crime that the police failed to prevent is to pursue a remedy under public law. None, however, is an effective alternative to a tort claim. Judicial review does not provide compensation for past wrongs. Victims of violent crimes can seek compensation from the Criminal Injuries Compensation Scheme, but the relevant awards are modest in comparison to tort damages, have a shorter limitation period of two years, and do not lead to an acknowledgment of the police's failings.Footnote 64 Another possibility is to pursue an express statutory remedy. The Police Reform Act 2002, amended by the Police Reform and Social Responsibility Act 2011, gives individuals the right to complain about the conduct of police officers.Footnote 65 The main shortcoming of this as an alternative remedy to a tort claim is that no compensation is payable as of right to individuals under it. A third possibility is to bring proceedings under another tort. The only other tort that can potentially apply here is misfeasance in public office. This is a general tort covering failures of all public authorities rather than a specialised regime for dealing with situations like the one discussed. As such, it is no more appropriate to apply than negligence. Therefore, the imposition of a duty of a care would not bypass the doctrinal restrictions of another more suitable tort.Footnote 66
4. Relationship with claims under the Human Rights Act 1998
The existence of the Human Rights Act (HRA) can potentially affect the negligence claim against the police for failure to prevent crime in two ways. First, it has been argued that the claim under the HRA operates as an alternative remedy, obviating the need to impose a duty of care at common law. Second, it has been suggested that the liability for pure omissions by public authorities is better dealt with under human rights law than in negligence.
(i) The HRA claim as an alternative remedy: In Osman v United Kingdom,Footnote 67 the European Court of Human Rights established that under the right to life in Article 2 of the European Convention on Human Rights, the police have a positive obligation to take preventive measures to protect individuals whose lives are at risk from the criminal acts of third parties. This means that a victim of crime who has suffered personal injury can bring an action under the HRA, by virtue of ss. 6 and 7, against the police for failure to prevent the crime. For this to succeed, the claimant needs to prove: (i) that the police knew or ought to have known at the time, (ii) that there was a real and immediate risk to the life of the victim of violence,Footnote 68 and (iii) that the police failed to take reasonable measures to avoid the risk. It could therefore be argued that the existence of a potential claim under the HRA has removed the need to impose a common law duty of care.Footnote 69
In our view this is unconvincing. First, although a claim under the HRA may result in damages, it has several disadvantages when compared to a claim under the tort of negligence.Footnote 70 Damages are not as of right, they tend to be lower than those in tort, and their assessment lacks clear guidance from the Strasbourg court.Footnote 71 Moreover, the limitation period for a HRA claim is only one year as compared with six years for tort. Second, as Nolan states, “the alternative remedy argument collapses if the protection offered by the Convention is not co-extensive with the law of negligence at a substantive level”.Footnote 72 The Osman test sets a high threshold for the establishment of liability and is very difficult to satisfy.Footnote 73 As such, it would not always be co-extensive with the scope of negligence liability.Footnote 74
(ii) The HRA claim as a better route for dealing with police omissions: Nolan has argued that the liability of the police for pure omissions is better treated as a matter of human rights law. He offers two main arguments in support. The first is that “the distinction between acts and omissions is foundational to the law of negligence … and the undermining of that distinction may therefore be expected to produce a degree of incoherence”.Footnote 75 The second is that moving away from the current position whereby public authorities obtain the benefit of the omissions principle, as private individuals do, would introduce an “alien public/private distinction” into private law by requiring the courts to distinguish between public authorities and private individuals in cases where the alleged duty is an affirmative one. By contrast, the distinction between positive and negative obligations is not as stark in human rights law.Footnote 76
The first argument in effect relies upon the claim that private law does not, and ought not to, recognise the existence of legal rights that others save one from suffering physical injury. We have already observed that the normative argument for this claim is weak. Indeed, it seems, on the contrary, that the law would be more coherent – if by that we mean more consistent with its underlying normative justifications – if the acts/omissions distinction were construed less rigidly in relation to public authorities. As we have seen, the arguments offered for the significance of that distinction do not withstand much scrutiny in respect of public authorities.
The second argument is also problematic. First, if we accept that the reasons offered for the omissions principle within private law apply less strongly to public authorities, then the distinction between public authorities and private individuals is one itself licensed by private law. Second, it need not follow from our argument that there is a strict distinction between public authorities and private individuals in relation to omissions liability. It may be that our arguments, such as the argument from dependence, do not apply as strongly to other public authorities. Third, the idea that a person should be subject to more extensive duties in virtue of that person's status is not at all alien to tort law. For instance, occupiers of land owe a duty of care to take positive steps to quell sources of danger which arise on the occupied land, even if such dangers were not created by their acts.Footnote 77
5. Fear of speculative or excessive litigation
The police, like other public authorities, are easy to trace and guaranteed to have financial resources to satisfy a claim. In addition, the third party inflicting the harm on the victim will often not be worth suing and, even if she is, she will almost always have less financial might than the police. In such cases, the rules on joint and several liability, which allow full recovery irrespective of the defendant's relative degree of culpability, mean that a claim against the police is more appealing.Footnote 78 This puts the police “at risk of speculative litigation in a way which most individuals and private-sector companies are not”.Footnote 79
This, however, does not lead to the conclusion that the police should never have a duty of care to prevent crime. What the argument shows is the need for adequate control devices when assessing the negligence liability of the police in such circumstances. As we explain in section V, even if one does not apply the omissions principle to the police, the existing framework of negligence liability regarding public authorities still provides adequate control devices.
6. Constitutional and institutional competence of courts
Negligence actions against public authorities, including the police, have the potential to engage the courts with questions of politics, such as the allocation of public funds and the prioritisation of competing citizens’ interests. The courts have traditionally been wary of intruding into such matters. This is broadly based on two grounds.Footnote 80 First, in a democracy, unelected judges should not overturn decisions that reflect the will of the electorate. Ultimately, it is for the democratic process to determine how public resources should be allocated and how conflicting values should be prioritised. Second, judges should be slow to make decisions on issues outside their technical competence which other bodies are better suited to determine.
These concerns do not militate against the imposition of a duty of care in every case of police failure to prevent crime. Instead, they suggest that the framework of negligence liability should include within it a mechanism for preventing the courts from determining claims which are not suitable for judicial resolution.
IV. The Rationale for a Duty of Care
In this section, we outline positive reasons for the imposition of a duty of care on the police in cases of failure to prevent crime. There are three central lines of argument. First, the police, like everyone else, owe a general moral duty to take reasonable steps to prevent reasonably foreseeable, avoidable physical suffering. This duty is correlated with a moral right, at least some of the time. Whilst it may not be appropriate legally to enforce everyone's moral duty to take reasonable steps to prevent avoidable suffering, the reasons against legal enforcement in relation to public authorities are, as already seen, weak. The upshot is that, by failing to take reasonable steps to prevent personal injury to individuals, the police breach their moral duty – a duty which there are no strong reasons not to enforce. Second, the police, additionally, owe a special moral duty in virtue of their special status. Third, the imposition of liability improves accountability.
A. “Wrongs Should Be Remedied”
In our view, the law should strive to ensure that certain categories of individuals suffering physical harm as a result of a wrongful failure by the police do not remain uncompensated. This reflects the much-cited dictum that “the rule of public policy which has first claim on the loyalty of the law” is that “wrongs should be remedied”.Footnote 81
The dictum of course begs the all-important question of what the law recognises as a wrong.Footnote 82 Clearly, the “wrong” in the dictum cannot be a legal wrong, since that is the precise issue to be resolved. However, this does not make the dictum meaningless. According to Lord Dyson, its meaning in practice is clear: “… prima facie if A foreseeably suffers harm as a result of the careless acts of B and there is a relationship of sufficient proximity between the two of them, then A should be compensated by B for the harm he has caused.”Footnote 83
The dictum's theoretical foundations have been defended by Robertson, who has argued that the dictum:
… seems to assume that the infringement of a duty that satisfies the first and second stages of the Caparo test constitutes a particular kind of moral wrong, the rectification of which is in the public interest. The first and second stages of the Caparo framework, then, are concerned with the question whether, as a matter of interpersonal justice (or interpersonal morality, the morality of what one person owes to another person, more broadly), the defendant should be regarded as owing a duty of care to the claimant.Footnote 84
This is different from saying that “the law should remedy harm caused by others”. The law is concerned with remedying wrongs, which “appear to be interpersonal moral wrongs for which the offender must pay, or infringements of duties that are justified by the notion of interpersonal responsibility identified through the elements of foreseeability and proximity”.Footnote 85
But is it possible for A to commit an interpersonal moral wrong against B by pure omission? Some deny it.Footnote 86 On these views, if we ever have moral obligations to save someone's life or to save them from serious harm (moral obligations arising absent some special relationship between two individuals), we never owe these obligations to the person whose life we morally ought to save. In short, none of our positive moral obligations is interpersonal in character.
Consider the strange consequences of this view. Virtually everyone agrees that, if A encounters a drowning child, B, in a shallow pool of water and A could, at almost no cost to herself, prevent B's death, and A knows all these facts, A has a moral obligation to do something.Footnote 87A idly stands by and breaches this obligation. A has done something wrong, but has A wronged the child? For the deniers, the answer is no. Yet it seems obvious to us that A has not simply behaved wrongly in some impersonal way, like the person who burns a beautiful painting which no one has or ever will see, but has behaved wrongfully in relation to the child. One argument for this follows from an interest theory of rights: A has paid insufficient regard to an extremely weighty interest of B's, namely B's interest in life; it is this failure to give adequate attention to B's interests, not “human interests in general”, which means that A wrongs B.
It might be objected that the fact that an interest theory of rights generates this conclusion is insufficient, since the most plausible moral theory of the legal rights in tort law is that the rights in tort law are rights which protect people's choices about how to use their bodies or property.Footnote 88 Such theories entail that A does not violate B's rights in the above example. If A fails to rescue the drowning child, A does not make a choice about the use of B's body which was B's to make.
However, as McBride has argued, the most plausible view is that, whilst some of our legal rights in tort law are indeed designed to protect certain choices which properly fall to us to make, this is not true of all of our legal rights in tort law.Footnote 89 Most pertinently, it is awkward to explain negligence liability as, in general, choice-protecting. The most natural explanation of why negligently running someone over is morally wrongful is that it substantially sets back their (non-choice-based) interests, not that it takes a decision which was the victim's to make. Therefore, from tort law's perspective, some of our rights are grounded in our interests.
If we accept, as tort law does, that a duty can be correlated with a right where that duty is grounded (primarily) in another person's interests, then we can accept that the moral duty to take steps to save the drowning child is correlated with a moral right of the child's: that duty is grounded primarily in the child's interest in life. In our view, it is the fact that the police have a similar moral duty to the victim which partly justifies the imposition of a legal duty to take care to prevent reasonably foreseeable physical injury in certain circumstances.Footnote 90
This raises some obvious objections. First, one might object that the argument proves too much: it would implausibly entail the recognition of legal duties on private individuals too. The premise of this objection is correct: the moral duty described is owed by everyone in A's position. However, it does not follow that it would be appropriate legally to enforce that moral duty against everyone in A's position. Some of the arguments in section II show that this would be inappropriate as against private individuals. But that section established that those objections do not apply to public authorities. So the prima facie case for enforcing the powerful moral duty to take affirmative steps to prevent serious physical injury for someone in A's position is not defeated by any countervailing considerations in the case of the police: wrongs should be remedied.
A second objection is that, if the ground for A's moral duty to take steps to assist B is B's important interest in avoiding death or serious injury, then this duty cannot coherently be limited to those who are in proximity to B. Consider the following argument from Epstein:
… even if the rule [imposing a duty to give positive assistance] starts out with … modest ambitions, it is difficult to confine it … X as a representative of a private charity asks you for $10 in order to save the life of some starving child in a country ravaged by war …. The money means “nothing” to you. Are you under a legal obligation to give the $10?Footnote 91
The concern is that, if the moral justification for the police's duty of care is, by its nature, so expansive, then any attempt to limit it in law will be morally arbitrary.Footnote 92
This is a serious concern to which we offer four responses. First, if it is partly the special powers of the police which permit legal enforcement of their moral duty, a natural limit to this enforcement is provided by the scope of these legal powers. Therefore, it may only be legitimate to enforce the moral duty we describe in situations concerning the prevention of crime within the territory covered by those special powers. Second, subject to the last point, our conception of proximity does not discriminate between victims based upon physical distance alone.Footnote 93 Third, although our conception of proximity differentiates between those at special risk of physical harm and those not so at risk, this distinction is not morally arbitrary. If the police know or ought to know that an individual is at special risk of personal harm, there will typically be identifiable, determinate, steps which can be taken to protect that individual. Such steps will be more difficult to identify where there is no such special risk.Footnote 94 Fourth, to the extent that our conception of proximity is not reflective of a difference in interpersonal morality, it needs to be borne in mind that considerations of practicality and administrative competence might justify limitations on the scope of legal rights which are not reflected in raw morality.Footnote 95 The law can only track morality so far.
B. Police's Special Status and Victims’ DependenceFootnote 96
The police's special status derives from the fact that they are “the specialist repositories for the state's monopolisation of legitimate force in its territory”.Footnote 97 The police are singled out in the broader interests of society as the primary body that are legally entitled to intervene and use force to protect citizens from criminal activity.Footnote 98 For that purpose, they are provided with specialist resources, equipment, and training. A person faced with the threat of violence is permitted by law to take reasonable measures of self-protection, but beyond that her only option is to inform the police.Footnote 99 In essence, other than reasonably protecting herself, the law obliges her to entrust her physical safety in the police. This gives rise to a relationship of dependence between the police and any member of the public who finds herself in that position.Footnote 100 Added to this is the fact that, in the majority of cases, a victim cannot protect or be reasonably expected to protect herself against the threat of violence. This further enhances the special status of the police and the relationship of dependence, the combination of which militates in favour of recognising that the police may potentially be liable in negligence in such cases.Footnote 101
It might be objected that private individuals are entirely at liberty to contract for personal security services to protect themselves against criminal conduct and so the premise of our argument – that the law requires us to enter into a relationship of dependence with the police – is false. There are two responses to this. First, some protective measures have been legally foreclosed to us by the state, namely bearing arms. In a world in which people were free to bear arms, who knows whether a particular victim would have been armed and whether this would have prevented her death. Given that this uncertainty has been created by the state, it is arguable that it should be borne by the state. Second, by setting up the police force, the state has adversely affected free competition in the marketplace for security services, with the result that the cost of personal security is probably higher than it otherwise would be – prohibitively expensive for many victims.
C. Public Accountability for Police Conduct
One of the functions that tort law performs when it applies to public officials is holding them to account.Footnote 102 The same is true of an action against the police for negligent investigation. Such an action forces the police to explain their conduct in respect of the facts relating to the injury in public and as part of the adversarial process.Footnote 103 This is important to victims and their relatives. Often, the real reason why these sue “is the desire for a proper investigation into what went wrong, with the possibility of a public condemnation at the end”.Footnote 104 In this respect, the tort litigation performs a vindicatory role too, which is “to provide a public and impartial public forum for declaring the claimant's rights (the legal origin of which must be established on other grounds) have been infringed”.Footnote 105
There are of course other ways by which the police can be held accountable in such cases.Footnote 106 These, however, do not undermine the accountability and vindicatory role of the tort action.Footnote 107 To begin with, the complaints procedure set out by the relevant legislation suffers from limitations that dent the public's confidence in it. Most of the complaints are dealt with internally by the relevant police force with the result that the process lacks the appearance of independence that judicial scrutiny has. Although it is possible for the Independent Police Complaints Commission to intervene in some cases, the effectiveness of this body remains questionable.Footnote 108 Overall, as the editors of a specialist work observe, “the chances of an aggrieved person being vindicated through a complaint are considerably lower than his chances of succeeding in a civil action”.Footnote 109 External public inquiries do not suffer from similar shortcomings, but they are time-consuming and costly and thus difficult to establish as frequently as it is required. Moreover, such inquiries cannot be set in motion by the individuals concerned.
V. The Framework of Negligence Liability
In this section, we argue that, when the existing framework of negligence liability is applied in light of the arguments in sections III and IV, it follows that a duty of care should be owed by the police in respect of pure omissions in certain types of case, the nature of which we describe.
A. Justiciability
The first question is whether the claim is justiciable, namely appropriate for judicial resolution. This is a preliminary hurdle, so that, if the answer is negative, the court will not inquire into the existence of a duty of care. Justiciability serves a useful function by ensuring that the courts do not decide cases outside their institutional competence.
In Connor v Surrey County Council,Footnote 110 Laws L.J. summarised the current approach as follows: (i) where a public authority makes “a pure choice of policy under a statute which provides for such a choice to be made”, the claim will be non-justiciable; (ii) where the decisions involve policy and operations, the court's conclusion would be sensitive to the particular facts, though “the greater the element of policy involved … the more likely it is that the matter is not justiciable”; (iii) where the decision is purely operational, it will be justiciable. This acknowledges that unclear cases are bound to arise. Nonetheless, given that justiciability is merely the first of several analytical steps in determining whether a public authority is liable in negligence, it is important not to have an excessively wide concept of it.Footnote 111
In the police context, the application of the test means that the deployment of resources and selection of strategy in an investigation that Lord Keith discussed in Hill are likely to be non-justiciable, whereas operational negligence by an officer would be justiciable. Rigby v Chief Constable of Northamptonshire Footnote 112 provides a good illustration. Firing the canister without a fire service in attendance was an operational negligence for which there could be liability, but the failure to obtain equipment delivering CS gas less dangerously was non-justiciable because the decision of what equipment to purchase was a policy one connected with the allocation of resources. Likewise, if the failure to respond to an emergency call is due to a police officer's individual act of carelessness, such as the call taker's missing of the call as a result of listening to music, the issue is justiciable. However, if it is due to an insufficient number of patrol cars arising from the allocation of resources, the issue would be non-justiciable.Footnote 113
B. Duty of Care
If the claim is justiciable, the court will consider whether a duty of care is owed. As Lord Bingham said in Van Colle/Smith:
… the most favoured test of liability is the three-fold test laid down by the House [of Lords] in Caparo Industries plc v Dickman [1990] 2 AC 605, by which it must be shown that harm to B was a reasonably foreseeable consequence of what A did or failed to so, that the relationship of A and B was one of sufficient proximity, and that in all the circumstances it is fair, just and reasonable to impose a duty of care on A towards B.Footnote 114
1. Reasonable foreseeability of harm
A duty of care will only be imposed where a reasonable person in the position of the defendant would have realised that her carelessness may cause the claimant to suffer the type of harm that she has suffered. Whether the requirement is satisfied depends on the facts of each case.
2. Proximity
“Proximity” denotes closeness of some sort between the parties at the time of the alleged negligence, but its precise meaning remains elusive.Footnote 115 The question here is how “proximity” can be established with a view to imposing an affirmative duty on the police to protect the claimant from harm by a third party.
Under the existing law, there are five possible categories. The first is where the police create a source of danger with which the third party interferes to cause harm.Footnote 116 The second is where the police's involvement prevents alternative means of rescue, thus making the situation worse.Footnote 117 The third is where the police have sufficient control over the third party who causes the damage to the claimant, such as if that party is a detainee.Footnote 118
The fourth category is where the police have “assumed responsibility” for the claimant's safety.Footnote 119 The phrase is hard to pin down,Footnote 120 but here it seems to require an undertaking by words or conduct to protect the claimant, coupled with reliance by the claimant on that. In Michael, Joanna Michael made an emergency call from her mobile phone to the police to inform them that her ex-partner had threatened to return to her house and kill her.Footnote 121 Although she lived in the area of South Wales Police, her call was received by a call handler for Gwent Police. The latter told Joanna that she would pass on the call to South Wales Police and that she should keep her phone free as South Wales Police would want to call her back. In forwarding the call to South Wales Police, the call handler seemingly failed to pass on that Joanna said that her ex-partner had threatened to kill her. As a result, the South Wales Police operator graded her call at a lower level of priority, leading to a slower response time. She had been murdered by the time the police eventually arrived at her house. The Court of Appeal ruled that there was no assumption of responsibilityFootnote 122 and the majority of the Supreme Court agreed. According to Lord Toulson, “the only assurance which the call handler gave to Ms Michael was that she would pass on the call to the South Wales Police. She gave no promise how quickly they would respond”.Footnote 123 This shows that the mere receipt of an emergency call by the police will not trigger an assumption of responsibility.Footnote 124 A promise or assurance of some kind, on which the claimant relies, is required.
The last category is where the claimant has a particular relationshipFootnote 125 with the police, which the courts have treated as entitling her to rely on protection by the police, such as if she is a detaineeFootnote 126 or an employee.Footnote 127 As it currently stands, this is of narrow application.
From the above analysis, it follows that, in many cases where a claimant makes an emergency call to the police to report a threat of violence, like in Michael, or reports such threats and related incidents to the police over a long period of time, like in Van Colle/Smith, there will be no proximity. In our view, this should be overturned. In particular, a finding of proximity should arise where the following factors are satisfied:
(1) The claimant is at a special risk of personal harm, namely a greater risk than the general public. The circumstances in which the risk will be special must be left to the courts to develop on a case-by-case basis. Guidance on this can be found in Couch v Attorney-General, where a majority in the New Zealand Supreme Court held that “the necessary risk must be … special in the sense that the plaintiff's individual circumstances, or her membership of the necessary class rendered her particularly vulnerable to suffering harm of the relevant kind” from the third party.Footnote 128 In any case, a person facing a specific threat to her physical safety from a specific individual is at a special risk.
(2) The police are aware or should have reasonably been aware that the claimant is at a special risk of personal harm.Footnote 129
(3) The police, by virtue of their special status, have the power to protect the class of persons to which the claimant belongs, namely members of the public at a special risk of personal harm.
(4) The claimant is dependent upon the police as regards protection against the risk on the basis of the legal and civic duties imposed on her to inform the police about the incident and to refrain from taking measures beyond reasonable self-protection.
According to this analysis, proximity is based on two broad elements: first, the distinguishability of a claimant from the general public and, second, her relationship of dependence with the police which is inextricably intertwined with the police's special status in this context. The test does not require a specific undertaking by the police to protect the claimant or a conscious reliance by the claimant on that. It would therefore cover a case where a police call handler simply does not answer a victim's emergency call because she is listening to music.
In terms of the analytical framework, the finding of proximity in these cases can be explained as an extension of the fifth category identified above. In our view, it is helpful to describe this fifth category as “status-based”Footnote 130 for two reasons. First, it assists in making clear that a duty of care may arise even in the absence of anything describable as a specific undertaking or representation made by the police to the claimant. Where a police call handler fails to answer a call because she is listening to music, it is entirely artificial to describe any duty of care as arising from an undertaking or representation made by her to the claimant. Second, this label captures the fact that it is by virtue of the special, dependence-creating status of the police that the claimant's being at special risk of harm can generate a relationship of proximity.
In Michael, the majority rejected Lady Hale's view that proximity can exist “if the police know or ought to know of an imminent threat of death or personal injury to a particular individual which they have the means to prevent”.Footnote 131 Three reasons were given: (i) the definition wrongly implies that a duty should be owed to “the intended victim of a drive-by shooting but not to an injured bystander”, (ii) the requirement that the threat be imminent is arbitrary, and (iii) the limitation of the duty to personal injury is arbitrary.Footnote 132
The criticisms do not apply, at least with equal force, to our formulation. The situation envisaged by (i) does not pose a difficulty. If the police know that a particular individual is at risk of being injured in a drive-by shooting at a particular location, it will also be true that people at that location will be at special risk of injury. They would therefore fall within our definition of proximity. We agree that (ii) is arbitrary.Footnote 133 One way in which a person may be at “special” risk of injury is where they face an imminent threat of injury, but an individual may nonetheless be at special risk where a threat is not imminent. The lack of imminence will be relevant, however, to what measures are required of the police in the circumstances and thus to whether they have breached their duty of care. Finally, the restriction to personal injury can be justified. We argued in section II that the justifications for the omissions principle do not preclude the existence of a duty of care upon the police. However, we were careful not to argue that the act/omission distinction is without moral significance. If that is so, then it is plausible that the extent of the positive obligations of care owed to individuals do not encompass all aspects of their wellbeing (unlike duties in respect of actions), but only the most important, such as personal injury.Footnote 134
3. Fairness, justice and reasonableness
At this stage, the courts balance policy factors for and against liability with a view to determining in which direction the law should incrementally develop.Footnote 135 This is followed in cases involving negligence claims against the police for failure to prevent crime.Footnote 136 The starting point in such cases was traditionally that a duty should not be imposed because of the policy factors adduced in Hill. We argued in section III that these should carry no weight. Instead, the courts should consider whatever valid policy factors arise on the facts of a case, recognising that one consideration, which is of utmost importance and must always be examined, is whether the imposition of a duty of care on the police would be compatible with the legal framework which sets out its powers and responsibilities. If the answer is no, then in our view there should be no duty of care. There are two reasons for this.
The first is deference to the will of the legislature.Footnote 137 Thus, where the framework is statutory and the statute explicitly provides that there should be no private law liability for the breach of a duty or the failure to exercise a power conferred by the statute on the public authority, it would be inappropriate to undercut this by imposing a duty of care in negligence. Most frequently, however, statutes are silent on whether they are actionable in private law. In Gorringe, Lord Scott stated that:
If the policy of the statute is not consistent with the creation of a statutory liability to pay compensation for damage caused by a breach of the statutory duty, the same policy would… 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.Footnote 138
Nonetheless, this analysis is not necessarily a correct interpretation of Parliament's intention in these cases. Parliament may well have intended to leave questions of liability to the common law, without attempting either to add a statutory liability to it or restrict it. Accordingly, Lord Steyn's approach in Gorringe, under which one asks whether the relevant statute excludes a private law remedy,Footnote 139 is preferable. This may allow a duty of care to arise at common law on the basis of the nature and purpose of the statutory functions without undermining Parliament's intention.
The second reason is to prevent the employees of a public authority from acting in a way that may undermine the primary purpose for which the statute or the common law has conferred powers on them. The courts must carefully examine the main purpose of the public authority's duty and ask whether the imposition of a duty of care would be inconsistent with that purpose in the sense of giving rise to a conflict of interest between the claimant and the class of persons intended to be protected.Footnote 140
The application of the consideration to cases of careless failure to prevent crime by the police as a result of which a proximate victim has foreseeably suffered personal injury requires an in-depth examination of the legal framework establishing the police's powers and duties. Although this is now partly found in statute, such as the Police Act 1996, s. 29 and Sch. 4, modified by the Police Reform Act 2002, s. 83, its roots lie in the common law.Footnote 141 The police's powers and duties have been discussed in several cases. In Brooks, Lord Steyn explained that the primary function of the police is to preserve the Queen's peace, which means concentrating on the prevention of the commission of crime, the protection of life and property, and the apprehension of criminals.Footnote 142 In Glasbrook Bros Ltd. v Glamorgan County Council, the House of Lords held that the police have an absolute duty to take all steps which appear necessary to them or in general for keeping the peace, for preventing crime, and for protecting from criminal injury.Footnote 143
Behind the similarity in these approaches lie important differences, especially in the way that the duty to protect life is understood. Lord Steyn in Brooks thought that the duty of the police is owed to the public at large and not to an individual victim, since such a duty would be detrimental to the exercise of the public function. In contrast, the earlier case law seems to envision an entitlement by an individual victim to protection by the police. In a passage cited with approval by the House of Lords in Glasbrook,Footnote 144 Pickford L.J. said in Glamorgan Coal Co. Ltd. v Glamorganshire Standing Joint Committee: “… if one party to a dispute is threatened with violence by the other party he is entitled to protection from such violence whether his contention in the dispute be right or wrong.”Footnote 145 The court in Glasbrook acknowledged that the police's duty is subject to limits. Lord Blanesburgh observed that the police's “absolute duty to afford protection to life and property was only … limited by the extent of their available resources and by the urgency of competing claims upon their services”.Footnote 146 The limitation, however, is not the same as that suggested by Lord Steyn in Brooks. It seems to require an inquiry into whether in the specific circumstances there were other more demanding cases to deal with, or whether committing adequate resources would put the police in a position where they would be unable to deal with other cases that may potentially arise. It suggests that the police should not protect an individual if attempting to do so will leave the public unprotected, but it does not support the view that in general the police have no duty to protect an individual because doing so would undermine the protection they offer to the public at large. Overall, the duties of the police as traditionally understood by the courtsFootnote 147 go beyond what Lord Steyn envisaged in Brooks. Modern legislation has followed nineteenth-century legislation in fortifying the common law position, thus little turns on that.
Against this background, the imposition of a duty of care in the circumstances identified here would be compatible with the legal framework establishing the powers and duties of the police. First, following Lord Steyn's approach in Gorringe, the relevant statutes do not exclude a private law remedy. Second, finding a duty of care in such cases would not require the police to act in a manner that undermines their primary purpose derived from the legal framework within which they operate. As Wilberg states:
… the primary duty of the police requires them to apprehend suspects and to protect life and property then a duty of care owed to a victim will not run directly counter to that primary duty nor discourage the discharge of that duty. To the contrary, that primary duty will be reinforced by a duty to take care to protect potential victims.Footnote 148
Finally, the imposition of a duty of care is in line with the traditional understanding of the legal framework establishing the functions of the police, which envisions a duty on behalf of the police to protect identifiable victims from personal injury as long as they have the resources to do so and that does not impinge on the safety of other members of the public.
C. Breach of Duty
The police would breach their duty only if they do not do something that the reasonable police officer would have done in those circumstances. Where police officers exercise “professional” judgment, as they do when tackling crime, the relevant test would be the Bolam test. According to this, a police officer will not be “guilty of negligence if he has acted in accordance with a practice accepted as proper by a responsible body” of professional men “skilled in that particular art”.Footnote 149 Therefore, a police officer can escape liability by showing that there is a responsible body of opinionFootnote 150 in the police, even if that is the minority, which support the decision she made. This makes it “difficult to substantiate a case of fault against the background of a variety of professional practices”.Footnote 151 In effect, the test ensures that police officers investigating and suppressing crime are not held liable in negligence too readily, even if they owe a duty of care to the claimant.Footnote 152
Furthermore, the resources available to the police are taken into account in the standard of care expected of them when performing services that they are obliged to provide, such as fighting crime.Footnote 153 This is because the rationale for the general rule of ignoring the defendant's resources when assessing the standard of care, such as that a person who does not have the necessary resources to carry out an activity safely should not elect to carry it out, does not apply in these cases.
D. Causation and Remoteness of Damage
The normal rules concerning causation and remoteness of damage apply. Thus the claimant will need to prove on the balance of probability that, had the police conformed to their duty of care, this would have prevented her injury. As ever, proving what would have happened rather than what did happen is not straightforward. Therefore, the factual causation element of the claim may often be a difficult hurdle.
VI. Conclusion
The article has advanced two propositions. The first is that the current law on the negligence liability of the police for failure to prevent a crime is unsatisfactory. The general rule of non-liability is based on two lines of argument, neither of which is persuasive. The application of the pure omission rule is problematic, whilst the Hill policy grounds do not stand up to close scrutiny. At the same time, the current law does not attach adequate significance to arguments in favour of liability. We have argued that there are two powerful sources of a duty of care: the fairness of enforcing a moral duty of rescue against the police and the dependence upon the police for protection from violence – a dependence which the state creates.
The second proposition is that the existing framework of negligence liability of public authorities can be readjusted to generate outcomes that better balance the valid considerations for and against liability in cases of police failure to prevent crime. The readjustment we propose suggests that the police should be liable where they fail to prevent a crime with the result that the claimant suffered personal harm by an act of a third party if the following conditions are satisfied:
(1) The failure relates to an issue which is justiciable.
(2) The police have a duty of care in respect of the failure. That will be so where:
(a) It is reasonably foreseeable that the police failure to prevent the crime will result in personal harm to the claimant.
(b) The claimant has a relationship of proximity with the police. In most cases, this will be established because: (i) the police have specifically assumed responsibility to her, or, (ii) she is at a special risk of harm, the police know or should have known about it, they have the power to protect her from the risk and she is in a position of dependence on them.
(c) The imposition of the duty of care is consistent with the performance by the police of their legal functions and there are no other valid policy considerations which on balance negate such imposition.
(3) The police breached their duty in the sense that their action or inaction fell below the standard of care expected of a professional police officer.
(4) The breach of their duty caused in a factual and legal sense the claimant's personal harm.
