Michael and others v The Chief Constable of South Wales Police and another

This appeal arises from the tragic murder of Joanna Michael by a former partner, which might have been prevented if the police had responded promptly to a 999 call made by Ms Michael. As I explain below, two police forces were involved, Gwent Police and South Wales Police, and there was a lack of effective liaison between them.

The claimants in the action are Ms Michael’s parents and her two young children. The defendants are the Chief Constables of Gwent Police and the South Wales Police. The claim is brought for damages for negligence at common law and under the provisions of the Fatal Accidents Act 1976 and Law Reform Miscellaneous Provisions Act 1934 (which I will refer to as the common law or negligence claim), and for damages under the Human Rights Act 1998 for breach of the defendants’ duties as public authorities to protect Ms Michael’s right to life under article 2 of the European Convention on Human Rights (which I will refer to as the human rights or article 2 claim). Originally there was also a claim for misfeasance in public office.

The police applied for the claims to be struck out or for summary judgment to be entered in their favour. At first instance His Honour Judge Jarman QC struck out by consent the claim for misfeasance in public office but in a carefully reasoned judgment he refused to strike out or give summary judgment on the negligence and article 2 claims. The Court of Appeal reversed Judge Jarman’s decision in part. They held unanimously that there should be summary judgment in favour of the defendants on the negligence claim for reasons given by Longmore LJ, with which Richards and Davis LJJ agreed. The majority upheld Judge Jarman’s decision that the article 2 claim should proceed to trial. Davis LJ dissented on that issue. He would have held that on the facts alleged by the claimants there was no possibility that the claim under article 2 could succeed.

The claimants appeal against the decision of the Court of Appeal on the negligence claim. The police cross appeal against the decision of the majority of the Court of Appeal on the article 2 claim. Since the court is considering as a matter of law whether the claims have a real possibility of success, it must be assumed for present purposes that all factual allegations made by the claimants are capable of being established. In relation to the negligence claim, the sole question is whether the police owed any duty of care to Ms Michael on the facts as they are alleged. If so, questions about whether there was a breach of duty and its consequences would be matters for the trial.

Ms Michael lived in Cardiff with her two children who were aged seven years and ten months at the date of her death. On 5 August 2009 at 2.29 am Ms Michael dialled 999 from her mobile phone. She lived in the area of the South Wales Police, but the call was picked up by a telephone mast in Gwent and was routed to the Gwent Police call centre. It was received by a civilian call handler. The conversation was recorded and it has been transcribed. Ms Michael said that her ex-boyfriend was aggressive, had just turned up at her house in the middle of the night and had hit her. He had found her with another man. He had taken her car to drive the other man home and had said that when he came back he was going to hit her. She said that he was going to be back “any minute literally”.

She was asked by the call handler if she could lock the doors to keep him out. She replied that she could lock the doors, but she did not know what he would do. She did not know if he had a key or how he got into her house.

The next part of the transcript reads:

There is no explanation on the face of the transcript why the last three words are preceded by “(inaudible)” and appear in square brackets; but according to the call handler, who later made a written statement after listening to the recording of the call, at several points there was interference and noise in the background. As to the words in question, she said:

The call ended with the call handler telling Ms Michael that her call had come through to Gwent Police and that she would pass the call on to the police in Cardiff. She added “they will want to call you back so please keep your phone free”.

The call was graded by Gwent Police as a “G1” call. This meant that it required an immediate response by police officers. Ms Michael’s home was no more than five or six minutes’ drive from the nearest police station.

The Gwent call handler immediately called South Wales Police and gave an abbreviated version of what Ms Michael had said. No mention was made of a threat to kill. South Wales Police graded the priority of the call as “G2”. This meant that officers assigned to the case should respond to the call within 60 minutes.

At 2.43 am Ms Michael again called 999. The call was again received by Gwent Police. Ms Michael was heard to scream and the line went dead.

South Wales Police were immediately informed. Police officers arrived at Ms Michael’s address at 2.51 am. They found that she had been brutally attacked. She had been stabbed many times and was dead. Her attacker was soon found and arrested. He subsequently pleaded guilty to murder and was sentenced to life imprisonment.

Data held by South Wales Police recorded a history of abuse or suspected domestic abuse towards Ms Michael by the same man. On four occasions between September 2007 and April 2009 incidents had been reported to the police and entries had been made on a public protection referral for domestic abuse form, but in two instances the risk indications section of the form was not completed.

The consequences are stark and tragic. Ms Michael has lost her life in the most violent fashion. Her children have lost their mother and breadwinner. Her parents have lost their daughter and have taken on the responsibility and work of bringing up their grandchildren.

An investigation by the Independent Police Complaints Commission led to a lengthy report. It contained serious criticisms of both police forces for individual and organisational failures.

The court received full written submissions from the appellants, the respondents and three interveners. Liberty and Refuge made joint written submissions. Separate but broadly similar written submissions were made by Cymorth i Ferched Cymru (Welsh Women’s Aid). The Court heard oral submissions on behalf of the appellants from Nicholas Bowen QC, on behalf of Liberty and Refuge from Karon Monaghan QC and on behalf of the respondents from Lord Pannick QC.

The arguments raised the following issues:

If the police are aware or ought reasonably to be aware of a threat to the life or physical safety of an identifiable person, or member of an identifiable small group, do the police owe to that person a duty under the law of negligence to take reasonable care for their safety?

Alternatively, if a member of the public (A) furnishes a police officer (B) with apparently credible evidence that a third party whose identity and whereabouts are known presents a specific and imminent threat to his life or physical safety, does B owe to A a duty to take reasonable steps to assess such threat and, if appropriate, take reasonable steps to prevent it being executed?

On the basis of what was said in the first 999 call, and the circumstances in which it was made, should the police be held to have assumed responsibility to take reasonable care for Ms Michael’s safety and therefore owed her a duty of care in negligence?

On the material before the Court, was there arguably a breach of article 2?

In order to set their legal arguments in context, the interveners and the appellants referred to a substantial body of material about the deep-rooted problem of domestic violence in our society, its prevalence and weaknesses which have been identified in the police response to it. According to official homicide statistics, since 2001 in the United Kingdom around 100 women have been killed every year by a current or former partner. A report published last year by Her Majesty’s Inspectorate of Constabulary made strong criticisms of the overall police response to victims of domestic abuse (Everyone’s Business: Improving the Police Response to Domestic Violence).

It was not suggested by anyone in this case that the law of negligence should be developed in a way which is gender specific, but it was submitted that the need to combat the evil of domestic violence should influence the development of the common law in relation to potential victims of violence generally. Ms Monaghan also relied on the United Kingdom’s international legal responsibilities.

The United Kingdom signed the Convention on the Elimination of All Forms of Discrimination Against Women (CEDAW) on 22 July 1981 and ratified it on 7 April 1986.

Article 2 of CEDAW imposes an obligation on states, among other things, to establish legal protection of the rights of women on an equal basis with men and to ensure through competent national tribunals and other public institutions the effective protection of women against any act of discrimination.

Discrimination is defined in article 1 as including any distinction, exclusion or restriction made on the basis of sex which has the effect of impairing the enjoyment by women of their human rights on a basis of equality of men and women. The Committee on the Elimination of Discrimination has issued a general recommendation on the subject of violence against women: General Recommendation No 19 (11^th session, 1992). It states that gender-based violence is a form of discrimination that seriously inhibits women’s ability to enjoy rights and freedoms on a basis of equality with men. It recommends, among other things, that state parties should ensure that laws against family violence and abuse give adequate protection to all women; that effective complaints procedures and remedies, including compensation, should be provided; and that measures that are necessary to overcome family violence should include civil remedies and criminal penalties where necessary in cases of domestic violence.

Civil remedies may of course take many forms. There is no specific recommendation that a victim of domestic violence should have a right to sue the police for damages in the case of domestic violence which could have been prevented by the police. Nor is the United Kingdom under an international legal obligation to provide a remedy in that form.

The United Kingdom has signed, but not yet ratified, the Convention on Preventing and Combating Violence Against Women and Domestic Violence (“Istanbul Convention”), which came into force on 1 August 2014. It requires, by article 4(1), state parties to “take the necessary legislative and other measures to promote and protect the right for everyone, particularly women, to live free from violence in both the public and the private sphere” and by article 5(2) to “take the necessary legislative and other measures to exercise due diligence to prevent, investigate, punish and provide reparation for acts of violence covered by the scope of this Convention that are perpetrated by non-state actors”.

Aside from the fact that this Convention has not yet been adopted by the United Kingdom, it leaves it to states to decide what measures are necessary to promote these objectives.

Ms Monaghan submitted that it is also highly arguable that gender equality has achieved the status of a peremptory norm (jus cogens) in international law within the meaning of article 53 of the Vienna Convention on the Law of Treaties (which defines a peremptory norm of general international law as “a norm accepted and recognised by the international community of states as a whole as a norm from which no derogation is permitted and which can be modified only by a subsequent norm of general international law having the same character”). There was no detailed argument on this point because on the assumption for present purposes that there is now a rule of customary international law which obliges states to prevent and respond to acts of violence against women with due diligence (as the Special Rapporteur on Violence Against Women concluded in a report dated 20 January 2006 to the Commission on Human Rights of the United Nations Economic and Social Council), it is a matter for individual states how they do so.

Ms Monaghan’s submission was more general. She submitted that the international documents added weight to the arguments in favour of adopting the interveners’ liability principle. Acceptance of that principle, it was submitted, would be an appropriate measure directed at preventing violence and remedying damage caused by the state’s failure adequately to address the problem.

It has been long established that the police owe a duty for the preservation of the Queen’s peace. The phrase has an old-fashioned sound but the principle remains true. Halsbury’s Laws of England, fifth ed (2013), Vol 84, para 40, states that the primary function of the constable remains, as in the 17^th century, the preservation of the Queen’s peace.

In Glasbrook Brothers Ltd v Glamorgan County Council [1925] AC 270 a colliery manager asked for police protection for his colliery during a strike. He wanted police officers to be billeted on the premises. The senior police officer for the area was willing to provide protection by a mobile force, but he refused to billet police officers at the colliery unless the manager agreed to pay for the additional service at a specified rate. The manager promised to do so, but when the police submitted their bill the company refused to pay it on the ground that it was the duty of the police to prove necessary police protection without payment. The police sued the colliery and won.

The House of Lords held that the police were bound to provide such protection as was necessary to prevent violence and to protect the mines from criminal injury without payment, but that it was lawful for the police to charge the colliery for extra protection, and that the judge had been entitled to find on the facts that the case fell into that category. Viscount Cave LC stated the nature of the duty of the police at pp 277-278:

To similar effect Lord Parker CJ said in Rice v Connolly [1996] 2 QB 414, p 419, that it is the duty of a police constable “to take all steps which appear to him necessary for keeping the peace, for preventing crime or for protecting property from criminal injury”.

The duty is one which any member of the public affected by a threat of breach of the peace, whether by violence to the person or violence to property, is entitled to call on the police to perform. In short, it is a duty owed to the public at large for the prevention of violence and disorder.

Under section 83 of the Police Reform Act 2002 (substituting Schedule 4 of the Police Act 1996) every constable is required to make the following attestation:

This reflects the common law duty of the police. In recent years the courts have considered on a number of occasions whether, and in what circumstances, the police may owe a private law duty to a member of the public at risk of violent crime in addition to their public law duty.

In Hill v Chief Constable of West Yorkshire [1989] AC 53 the claimant was the mother of the last victim of a notorious murderer. Between 1975 and 1980 he murdered 13 young women in West Yorkshire. The statement of claim alleged that the police made a number of mistakes in their investigation which should not have been made by a competent police force exercising reasonable care and skill. For the purpose of deciding whether Mrs Hill had a valid claim against the police in negligence, the House of Lords assumed that the factual allegations were true, and that if the police had exercised reasonable care the murderer would have been arrested before he had an opportunity to murder her daughter. It was held that the police were under no liability in negligence.

The leading speech was given by Lord Keith of Kinkel. He recognised that the general law of tort applies as much to the police as to anyone else. Examples of police liability for negligence were Knightley v Johns [1982] 1 WLR 349 (where a police officer who attended the scene of a road accident carelessly created an unnecessary danger to the claimant) and Rigby v Chief Constable of Northamptonshire [1985] 1 WLR 1242 (where a police officer attending a break-in to a gunsmith’s shop carelessly caused severe damage to the premises by the firing of a canister into the building in the absence of fire-fighting equipment). But he held that the general duty of the police to enforce the law did not carry with it a private law duty towards individual members of the public.

Counsel for Mrs Hill relied on Anns v Merton London Borough Council [1978] AC 728 as authority for the proposition that the police, having decided to investigate the Yorkshire murderer’s crimes, owed to his potential future victims a duty to do so with reasonable care. The foundation of the duty was said to be the foreseeability of harm to potential future victims if the murderer were not apprehended. This, it was submitted, was sufficient to give rise to a duty of care applying Lord Atkin’s statement of principle in Donoghue v Stevenson [1932] AC 562 and Lord Wilberforce’s two stage liability test in Anns. Lord Keith rejected the argument. He emphasised that foreseeability of harm was not itself a sufficient basis for a duty of care in negligence. Some further ingredient was needed to establish the requisite proximity of relationship between the claimant and the defendant, and all the circumstances of the case had to be considered and analysed in order to ascertain whether such an ingredient was present.

Lord Keith referred to the decision of the House of Lords in Dorset Yacht Co Ltd v Home Office [1970] AC 1004, where Lord Diplock said (at p 1058) that the development of the law of negligence proceeds by first identifying the relevant characteristics of the conduct and relationship between the parties involved in the particular case and the kinds of conduct and relationships which have been held in previous decisions to give rise to a duty of care. In that case it was held that an action in negligence could lie against prison officers who negligently allowed young offenders camping on an island under the prison officers’ supervision to escape from the island by stealing the plaintiffs’ yacht. The reason for imposing liability was that the prison officers were responsible for exercising proper control over the wrong-doers, who were in their charge, and there was sufficient proximity between the prison officers and the owners of yachts in the close vicinity of the camp, because the use of their property as a means of escape was the very thing which the prison officers ought to have foreseen. By bringing the young offenders onto the island and leaving them unsupervised, the prison officers created a danger for the owners of the yachts which would not otherwise have existed.

In contrast, Lord Diplock said (at p 1070) that the courts would be exceeding their function in developing the common law to meet changing conditions if they were to recognise a duty of care to prevent criminals escaping from custody owed to a wider category of members of the public than those whose property was exposed to “an exceptional added risk by the adoption of a custodial system for young offenders which increased the likelihood of their escape unless due care was taken”.

Lord Keith said that if no general duty of care was owed to individual members of the public to prevent the escape of a known criminal, there could not reasonably be imposed on the police a duty of care to identify and apprehend an unknown one. Ms Hill could not be regarded as a person at special risk because she was young and female. She was one of a vast number of the female general public at risk from the murderer’s activities. He concluded that there was no ingredient or characteristic giving rise to the necessary proximity between the police and Ms Hill, and that the circumstances of the case were not capable of establishing a duty of care owed towards her by the police.

If Lord Keith had stopped at that point, it is unlikely that the decision would have caused controversy. It is not suggested in the present case that the decision itself was wrong. If the interveners’ liability principle is correct, it would not have assisted Mrs Hill, because her daughter was not an identifiable victim or a member of an identifiable small group.

However, having observed that what he had said was sufficient for the disposal of the appeal, Lord Keith went on to discuss the application of the second stage of Lord Wilberforce’s two stage test in Anns, if there had been potential liability under the first stage (at p 63). He concluded that it would be contrary to the public interest to impose liability on the police for mistakes made in relation to their operations in the investigation and suppression of crime. He said that the manner and conduct of such an investigation must necessarily involve a variety of decisions to be made on matters of policy and discretion, such as which particular line of inquiry is most advantageously to be pursued and what is the most advantageous way to deploy available resources. Many such decisions would not be appropriate to be called in question, but elaborate investigation of the facts might be necessary to ascertain whether or not this was so. A great deal of police time and expense might have to be put into the preparation of a defence to the action. The result would be a significant diversion of police manpower and attention from their most important function. He also said that the imposition of liability might lead to the exercise of the investigative function being carried out in a defensive frame of mind. He concluded that the Court of Appeal had been right to take the view that the police were “immune from an action of this kind on grounds similar to those which in Rondel v Worsley [1969] 1 AC 191 were held to render a barrister immune from actions for negligence in his conduct of proceedings in court”.

An “immunity” is generally understood to be an exemption based on a defendant’s status from a liability imposed by the law on others, as in the case of sovereign immunity. Lord Keith’s use of the phrase was, with hindsight, not only unnecessary but unfortunate. It gave rise to misunderstanding, not least at Strasbourg. In Osman v United Kingdom (1998) 29 EHRR 245 the Strasbourg court held that the exclusion of liability in negligence in a case concerning acts or omissions of the police in the investigation and prevention of crime amounted to a restriction on access to the court in violation of article 6. This perception caused consternation to English lawyers. In Z v United Kingdom (2001) 34 EHRR 97 the Grand Chamber accepted that its reasoning on this issue in Osman was based on a misunderstanding of the law of negligence; and it acknowledged that it is not incompatible with article 6 for a court to determine on a summary application that a duty of care under the substantive law of negligence does not arise on an assumed state of facts.

In Brooks v Commissioner of Police of the Metropolis [2005] UKHL 24, [2005] 1 WLR 1495, the claimant and his friend Stephen Lawrence were set upon by a gang of white youths in a racist attack. Stephen Lawrence was murdered. The claimant was traumatised. He claimed that the police owed him a duty of care in negligence: (a) to take reasonable steps to assess whether he was a victim of crime and, if so, to accord him reasonably appropriate protection and support; (b) to take reasonable steps to afford him the protection, assistance and support commonly afforded to a key eye-witness to a serious crime of violence; and (c) to afford reasonable weight to the account given by him and to act on the account accordingly.

The House of Lords held that the police owed him no such legal duty of care. All the judges endorsed the correctness of the decision in Hill but they expressed reservations about the width of some of the observations in Hill (per Lord Bingham at para 3, Lord Nicholls of Birkenhead at para 6 and Lord Steyn at para 28). It is clear that the part of Lord Keith’s speech to which they were referring was the final part in which he addressed the second stage of Lord Wilberforce’s test in Anns.

Lord Steyn (with whom Lords Rodger of Earlsferry and Brown of Eaton-under-Heywood agreed) said that the principle in Hill should be reformulated in terms of the absence of a duty of care rather than a blanket immunity (para 27). He noted that it was conceded by the police that cases of assumption of responsibility under what he described as the extended Hedley Byrne doctrine (Hedley Byrne & Co Ltd v Heller & Partner Ltd [1964] AC 465) fall outside the principle in Hill (para 29).

However, he said that the core principle of Hill had remained unchallenged in domestic jurisprudence and European jurisprudence for many years, and that if a case such as Hill were to arise for fresh decision it would undoubtedly be decided in the same way. He reiterated that the prime function of the police is the preservation of the Queen’s peace. The police must concentrate on preventing the commission of crime; protecting life and property; and apprehending criminals and preserving evidence. He said that a retreat from the principle in Hill would have detrimental effects for law enforcement:

By endorsing the principle in Hill in the terms that he did, Lord Steyn confirmed that the functions of the police which he identified were public law duties and did not give rise to private law duties of care (whether to victims, witnesses or suspects), although this did not exclude liability under Hedley Byrne.

Lord Bingham and Lord Nicholls were also of the view that the public duties of the police would potentially be impeded by the imposition of the duties asserted by Mr Brooks. Lord Bingham said that the duties pleaded could not be imposed on police officers charged in the public interest with the investigation of a very serious crime without “potentially undermining the officers’ performance of their functions, effective performance of which serves an important public interest” (para 4). Lord Nicholls was of the same view that the three legal duties asserted by the claimant “would cut across the freedom of action the police ought to have when investigating serious crime” (para 5).

In Brooks Lord Steyn referred to an argument that Hill should be distinguished on the basis that in that case the police negligence was the indirect cause of Ms Hill’s murder whereas in Brooks the behaviour of the police was a direct cause of harm to him. Lord Steyn observed that this did not do justice to the essential reasoning in Hill and he described the distinction as unmeritorious (para 32).

In Van Colle v Chief Constable of the Hertfordshire Police and Smith v Chief Constable of Sussex Police [2009] AC 225 the House of Lords heard together two appeals involving in different ways the question formulated by Lord Bingham as follows: if the police are alerted to a threat that D may kill or inflict violence on V, and the police take no action to prevent that occurrence, and D does kill or inflict violence on V, may V or his relatives obtain civil redress against the police, and if so, how and in what circumstances?

In Van Colle threats were made against a prosecution witness in the weeks leading to a trial. They included two telephone calls from the accused to the witness. The second call was aggressive and threatening but contained no explicit death threat. The witness reported the threats to the police. The matter was not treated with urgency. An arrangement was made for the police to take a witness statement, after which the police intended to arrest the accused, but in the interval the witness was shot dead by the accused. His parents brought a claim against the police under the Human Rights Act 1998 relying on articles 2 and 8 of the Convention. There was no claim under common law. The police were held liable at first instance and failed in an appeal to the Court of Appeal, but succeeded in an appeal to the House of Lords.

The House of Lords applied the test laid down by the Strasbourg court in Osman (para 116) for determining when national authorities have a positive obligation under article 2 to take preventative measures to protect an individual whose life is at risk from the criminal acts of another:

The critical question of fact was whether the police, making a reasonable and informed judgment at the time, should have appreciated that there was a real and immediate risk to the life of the victim. The House of Lords held that the test was not met.

Smith reached the House of Lords on an application to strike out. The question was whether the police owed a duty of care to the claimant on the assumed facts. The claimant was a victim of violence by a former partner. He had suffered violence at the hands of the other man during their relationship. After it ended, he received a stream of violent, abusive and threatening messages, including death threats. He reported these matters to the police and told a police inspector that he thought that his life was in danger. A week later the man attacked the victim at his home address with a claw hammer, causing him fractures of the skull and brain damage. The assailant was subsequently convicted of making threats to kill and causing grievous bodily harm with intent. The House of Lords held by a majority that the police owed the victim no duty of care in negligence.

Lord Bingham, dissenting, formulated his liability principle which I have set out.

Lord Bingham’s starting point was that the circumstances in which A will be held liable in negligence for unintended harm suffered by B depend on the relationship between them. He recognised that it is not usual for A to be liable to B where harm is caused to B by a third party C, but said that in some circumstances A might be liable for such harm if A should have prevented C. In some cases A’s liability had been found to depend on an assumption of responsibility by A towards B; and in other cases, notably Dorset Yacht, on the finding of a special relationship between A and C by virtue of which A was responsible for controlling C. Currently, he said, the most favoured test of liability was the three-fold test laid down by the House of Lords in Caparo Industries Plc v Dickman [1990] 2 AC 605.

Lord Bingham did not consider that his liability principle conflicted with the ratio of either Hill or Brooks, or that it would distract the police from their primary function of suppressing crime and apprehending criminals. He observed that statements in Glasbrook Bros Ltd v Glamorgan County Council and Glamorgan Coal Co Ltd v Glamorganshire Standing Joint Committee [1916] 2 KB 206 (referring to protection of property) would support a broader liability principle, but he said that the law attached particular importance to the protection of life and physical safety, and he did not consider it necessary to analyse in detail the cases on property damage.

Lord Bingham did not consider that the policy reasons given by Lord Keith in Hill justified the width of what he said about police immunity.

Lord Hope of Craighead (with whom Lord Carswell and Lord Brown agreed) shared Lord Bingham’s view that the reasons given by Lord Keith in Hill for saying that an action for damages for negligence should not lie against the police on grounds of public policy did not all stand up to critical examination. He regarded Brooks as a more important authority. In disagreement with Lord Bingham, he considered that the risks identified in Brooks of imposing principles which would tend to inhibit a robust approach in addressing a person as a possible suspect or victim were relevant to cases of which Smith was an example.

Lord Hope recognised that Lord Bingham’s liability principle was confined to cases where a member of the public furnished apparently credible evidence to the police that a third party represented a specific and imminent threat to his life or physical safety, but he considered that this formulation would lead to uncertainty in its application and to the detrimental effects about which Lord Steyn had warned in Brooks.

Lord Phillips of Worth Matravers CJ identified the core principle in Hill as being that in the absence of special circumstances the police owe no common law duty of care to protect individuals against harm caused by criminals (para 97). The question was whether that core principle could stand with, or accommodate by way of exception, the liability principle formulated by Lord Bingham. As to that, he did not find it easy to identify the essential parameters of the principle. He asked rhetorically whether the principle would apply when the evidence emanates from a third person; or if the whereabouts but not the identity of the potential wrongdoer was known; or if the threat was specific, but not imminent; or if the threat was imminent but not specific. He also questioned why the principle should be restricted to a threat to life or physical safety, and not apply to a threat to property. He concluded (para 100) that the elements in Lord Bingham’s liability principle were facts which would make particularly egregious a breach of duty of care that could be more simply stated:

Hill, Brooks and Van Colle and Smith are the most important decisions but some others deserve mention.

In Calveley v Chief Constable of the Merseyside Police [1989] AC 1228 the House of Lords upheld a decision striking out claims in negligence for damages for lost overtime by police officers who had been suspended pending disciplinary proceedings which ended in their reinstatement. They alleged that they were owed a duty by the investigating officers to exercise proper care and expedition in the conduct of the investigation which had not been met. It was argued that a police officer investigating a suspected crime owes a duty of care to the suspect and that the same principle applied to the investigation of a disciplinary offence. The House of Lords rejected the argument, which Lord Bridge of Harwich described as startling (p 1238). He said that other considerations apart, it would be contrary to public policy to prejudice the fearless and efficient discharge by police officers of their vitally important public duty of investigating crime by requiring them to act under the shadow of a potential action for damages for negligence by the suspect.

Similarly in Elguzouli-Daf v Commissioner of Police of the Metropolis [1995] QB 335 the Court of Appeal upheld decisions striking out actions for negligence brought by claimants who had been arrested and held in custody during criminal investigations which were discontinued. Steyn LJ, in the leading judgment, added the qualification that there might be a case in which the Crown Prosecution Service assumed by its conduct a responsibility towards a particular defendant under the Hedley Byrne principle, as expounded by Lord Goff of Chieveley in Spring v Guardian Assurance Plc [1995] 2 AC 296.

In that case Lord Goff said that Hedley Byrne was widely regarded as a case on liability in damages for negligent misstatement and liability in negligence for economic loss, which it was, but that it was important not to lose sight of the underlying wider principle. The underlying principle rested on an assumption of responsibility by the defendant towards the plaintiff, coupled with reliance by the plaintiff on the exercise by the defendant of due skill and care. The principle that a duty of care could arise in that way was not limited to a case concerned with the giving of information and advice (Hedley Byrne) but could include the performance of other services.

Elguzuli-Daf was cited with approval in Brooks and in Van Colle and Smith.

An Informer v A Chief Constable [2013] QB 579 provides an example of a duty of care arising from an assumption of responsibility coupled with reliance by the claimant. The claimant contacted the police regarding the activities of a business associate. He was introduced to two police contact handlers. He agreed to act as an informant under the police instructions and he later signed a set of instructions prepared by the police. At the outset they explained the steps which they would take to protect his identity and gave him assurances that they would treat his safety and that of his family as a priority. As the investigation developed the claimant himself became a suspect. A restraint order was obtained against him under the Proceeds of Crime Act 2002, prohibiting him from disposing of his assets, but the Crown Prosecution Service eventually decided not to prosecute him and the restraint order was discharged. He sued the police, alleging that they owed him a duty of care to protect his economic interests. The police conceded that they owed a duty of care to protect his physical well-being, and that of his family. They had assured him that they would do so and he had acted on the faith of their assurances. But they had given him no assurances that they would protect his economic interests and the Court of Appeal upheld the judge’s decision that they owed him no such duty, which would potentially conflict with their responsibility to the public for the investigation of crime and the proceeds of crime.

There have been cases of a police force being held liable in negligence for failing to take proper care for the protection of a police officer against a criminal attack, but they were based on the duty of care owed to the claimants as employees whose employment exposed them to the risk of such an attack in the performance of their duty: Costello v Chief Constable of Northumbria [1999] ICR 752, Mullaney v Chief Constable of the West Midlands [2001] EWCA Civ 700.

Claims against other emergency services have been treated in a similar way to claims against the police (except in the case of the ambulance service, to which I refer below). In Capital & Counties Plc v Hampshire County Council [1997] QB 1004 the Court of Appeal considered claims in negligence against fire authorities arising out of three incidents in which the fire brigade responded to a 999 call.

In the first case the fire brigade was called to a fire at office premises in Hampshire. The fire triggered the operation of a heat-activated sprinkler system, but on arrival a fire brigade officer gave instructions for the sprinkler system to be shut down. This led to the fire rapidly spreading out of control and the premises were destroyed. If the sprinkler system had been left on and the fire brigade had otherwise acted as it did to combat the fire, the premises would not have been destroyed.

In the second case the fire brigade was called to the scene of some fires on waste land near to the claimants’ industrial premises in London. When the fire brigade arrived the fires had already been extinguished. After checking that there was no evidence of any continuing danger the fire brigade left. Later a fire broke out at the claimants’ premises. They sued the fire authority alleging negligence in failing properly to inspect the wasteland and failing to ensure that all fires and risk of further fires in the area had been eliminated before leaving.

In the third case the fire brigade was called to a fire at a chapel in Yorkshire. The water hydrants near the premises either failed to work or the officers were unable for a long time to locate them, and so water had to be fetched from a dam half a mile away. It should have been possible to contain the fire, but as a result of the water shortage the whole building was destroyed. Under the Fire Services Act 1947 fire authorities were under a statutory duty, among other things, to secure the services for their area of a fire brigade and equipment, such as necessary to meet efficiently all normal requirements, and to take all reasonable measures to ensure that an adequate supply of water was available for use in case of fire. The owners of the chapel sued the fire authority for negligence and breach of statutory duty. They alleged that there ought to have been a proper system of inspection to ensure that hydrants were in working order and that the fire crew were at fault in failing to locate some of the hydrants sooner.

The Court of Appeal upheld decisions to allow the claim in the Hampshire case but to dismiss the claims in the London and Yorkshire cases. The difference was that in the Hampshire case the fire brigade aggravated the situation by causing the sprinkler system to be turned off, whereas in the other cases the failures of the fire brigade made things no worse than they were. In drawing that distinction the court applied the reasoning of the House of Lords in Geddis v Proprietors of Bann Reservoir (1878) 3 App Cas 430, 455 (per Lord Blackburn) and East Suffolk Rivers Catchment Board v Kent [1941] AC 74. In the Hampshire case the court also drew an analogy with Rigby v Chief Constable of Northamptonshire and Knightley v Johns, where the rescue service created additional danger.

There are two Scottish decisions at first instance in fire brigade cases in which the Lord Ordinary took a different direction: Duff v Highlands and Islands Fire Board 1995 SLT 1362 and Burnett v Grampian Fire and Rescue Services 2007 SLT 61. In Duff the fire service attended a house fire and apparently extinguished it. After they left, fire broke out again destroying the property and the house next door. Lord Macfadyen dismissed the claim for want of proof of negligence, but he said that he would have rejected the defenders’ argument that they could not be held liable for damage which would have occurred if they had done nothing.

Duff was followed by Lord Macphail in Burnett. That was a similar case in which a fire re-ignited after the fire brigade had left. On a preliminary plea by the defenders to the relevancy of the pursuer’s averments, Lord Macphail declined to follow the reasoning of the Court of Appeal in Capital & Counties Plc v Hampshire County Council and ruled that the case should go to trial. He said that the law of Scotland does not draw a distinction between acts and omissions comparable to that which appeared to exist in the English law of tort, and that the decision in Capital & Counties Plc v Hampshire County Council did not represent the law of Scotland (paras 34 and 48).

Burnett was cited in Mitchell v Glasgow City Council [2009] AC 874, to which I refer below. Burnett was not mentioned in the judgments, but the distinction between acts or omissions was central to Lord Hope’s reasoning, and he observed that the law of liability for negligence has developed on common lines both north and south of the Border (para 25).

In OLL Ltd v Secretary of State for Transport [1997] 3 All ER 897 May J struck out claims against the Coastguard for negligence in responding to a 999 call. The Coastguard is a non-statutory public authority with responsibility for organising and coordinating search and rescue missions on the coast and at sea. The claims arose from a tragic incident in which a party of children and teachers, under the supervision of instructors from an adventure centre, got into difficulties on a canoeing trip. The Coastguard was alerted but several hours passed before all the members of the party were rescued. Some of the children died and others suffered severe hypothermia. It was alleged that the Coastguard was negligent in that it was slow to launch a search and rescue operation and misdirected a lifeboat and a helicopter about where they should search. In striking out the claims the judge applied the reasoning in Capital & Counties Plc v Hampshire County Council. He rejected the argument that the misdirection of searchers was analogous to the fire brigade turning off the sprinkler system in the Hampshire case, because it did not positively increase the danger to the canoeists. It was similar to a fire brigade sending one of its fire engines to the wrong address.

In Van Colle and Smith Lord Bingham reserved his opinion about the correctness of Capital & Counties Plc v Hampshire County Council and disapproved OLL Ltd v Secretary of State for Transport. But he was alone in criticising them, and he did so in the context of formulating his liability principle which the other members of the House of Lords rejected.

The position of the ambulance service was considered by the Court of Appeal in Kent v Griffiths [2001] QB 36. A doctor attended the home of a patient suffering from an asthma attack and called for an ambulance to take her immediately to hospital. The control replied “Okay doctor.” After 13 minutes the ambulance had not arrived and the patient’s husband made a further call. He was told that an ambulance was well on the way and should arrive in seven or eight minutes. For unexplained reasons it did not arrive until 40 minutes after the first call. The patient suffered a respiratory arrest which would have been prevented if the ambulance had arrived in a reasonable time. The patient’s doctor gave evidence that if she had been told that it would take the ambulance service 40 minutes to come, she would have advised the patient’s husband to drive her to hospital and would have gone with them. The Court of Appeal upheld the trial judge’s finding of liability against the ambulance service. It would have been sufficient to hold that the acceptance of the doctor’s request for an ambulance to come immediately gave rise to a duty of care but Lord Woolf MR (with whom the other members of the court agreed) went further. He held that the ambulance service, as part of the health service, should be regarded as providing services equivalent to those provided by hospitals, and not as providing services equivalent to those rendered by the police and fire services. Accordingly, the staff of the ambulance service owed a similar duty of care to that owed by doctors and nurses operating in the health service (para 45).

Courts in other common law jurisdictions have taken various approaches.

In the USA the matter is governed by the tort law of individual states. In New York the Court of Appeal has held, by a majority, that the police do not owe a duty of care in negligence for the protection of members of the public, unless they undertake a duty to protect particular members of the public and expose them without adequate protection to risks which materialise: Riss v City of New York 22 NY 2d 579, 240 NE 2d 860 (1968), distinguishing Schuster v City of New York 5 NY 2d 75, 180 NYS 2d 265, 154 NE 2d 534 (1958). Similarly, in the case of emergency calls, the position generally appears to be that the police will owe a duty of care only if the call handler gives an explicit assurance on which the caller relies: Cuffy v City of New York 69 NY 2d 255 (1987), Noakes v City of Seattle 77 Wash App 694, 895 P2d 842, 845 (1995), Perkins v City of Rochester 641 F Supp 2d 168 (2009).

In South Africa, the leading case is the decision of the Constitutional Court in Carmichele v Minister of Safety and Security (2001) 12 BHRC 60. The applicant was brutally attacked by a man awaiting trial for attempted rape. The police and prosecutor had recommended his release on bail despite a history of sexual violence. The applicant sued the ministers responsible for the police and prosecution service, alleging that they had negligently failed to see that the magistrate was properly informed about the risk he posed to women in the vicinity of his home, including the applicant. Her claim was dismissed by the High Court and its decision was upheld by the Supreme Court of Appeal, but she succeeded on appeal to the Constitutional Court, relying on a provision in section 39(2) of the constitution which required the courts when developing the common law to “promote the spirit, purport and objects of the Bill of Rights”. The Constitutional Court decided that it would not be appropriate for itself to determine whether the law of delict required to be developed so as to afford a right to the applicant to claim damages if the police or prosecutor were negligent. It said that it was by no means clear how the constitutional obligations on the state should translate into private law duties towards individuals, and that the court would be at a grave disadvantage in deciding the issue without a fully reasoned judgment of the High Court or Court of Appeal. It set aside the decisions of the lower courts and remitted the matter to the High Court. The discussion in the judgment is interesting, but the decision itself is of little help, not only because it left the matter undetermined but because it was based on the provisions of the South African constitution and Bill of Rights.

In Hamilton v Minister of Safety and Security [2003] 4 All SA 117 the Supreme Court of Appeal held the police liable to the victim of a shooting for negligently issuing a firearm licence to the attacker, who had a history of psychosis, personality disorder and alcohol abuse. The agreed statement of facts did not suggest that the victim was at higher risk than any other member of the public.

In Canada, the Divisional Court of the Ontario High Court refused an application to strike out a claim in negligence by the victim of a serial rapist against the police for their failure to warn potential victims living in the area about the risk which they faced: Jane Doe v Metropolitan Toronto (Municipality) Commissioners of Police 74 OR (2d) 225 (1990), 72 DLR (4^th) 580. In a short judgment the court applied Lord Wilberforce’s two stage test in Anns and concluded that the facts pleaded were sufficient to establish a special relationship of proximity. The claimant later succeeded at the trial: (1998) 160 DLR (4^th) 697.

In Hill v Hamilton-Wentworth Regional Services Board [2007] 3 SCR 129 a wrongly convicted defendant sued the police for negligent investigation of the case against him, alleging that he should never have been a suspect. The Supreme Court held, by a majority, that a duty of care existed between the police and a suspect in a criminal investigation. McLachlin CJ, giving the judgment of the majority, expressly limited the judgment to that relationship. She said that it might well be the case that the considerations informing the analysis of proximity and policy would be different in the case of the relationship between the police and a victim; and that if a new relationship was alleged to attract liability of the police in negligence in a future case, it would be necessary to engage in a fresh Anns analysis. She disclaimed reliance on Jane Doe v Metropolitan Toronto (Municipality) Commissioners of Police, describing it as of little help. She noted that it was a lower court decision and that debate continued over the content and scope of its ratio (para 27).

In New Zealand the highest authority is the decision of the Supreme Court in Couch v Attorney-General [2008] 3 NZLR 725. Victims and relatives of victims injured or killed in a robbery claimed damages in negligence for the alleged failure of the probation service to exercise reasonable care in the supervision of the offender, who was on licence after release from a prison sentence for aggravated robbery. The victims were employed at a club where the attacker had been allowed by the probation service to obtain work experience without the knowledge of the employer and his fellow employees about his background. The Supreme Court allowed an appeal by the claimant from the decision of the Court of Appeal that the claim should be struck out. Its decision was unanimous but there were differences as to the criteria for establishing a duty of care.

The reasoning of the majority (Blanchard, Tipping and McGrath JJ) was given by Tipping J. He took as his starting point the well-known observation of Dixon J in Smith v Leurs (1945) 70 CLR 256, 262 that it is exceptional to find a duty in law to control another’s actions to prevent harm to strangers, but that special relations may be the source of a duty of this nature. Tipping J noted that the special relations to which Dixon J referred were between the defendant and the wrongdoer, but there had additionally to be a special relationship between the defendant and the claimant - special in the sense that there was sufficient proximity between the parties to make it fair, just and reasonable, subject to matters of policy, to impose the duty of care in issue (para 85).

Tipping J concluded that the power of the probation board over the wrongdoer’s employment was arguably sufficient to establish the necessary relationship between the defendant and the wrongdoer, by analogy with the Dorset Yacht case. As to the relationship between the defendant and the claimant, the necessary proximity criterion would be satisfied if she could show (as was arguable on the facts) that she, as an individual or a member of an identifiable and sufficiently delineated class, was the subject of a distinct and special risk of suffering harm. The necessary risk must be distinct in the sense of being clearly apparent, and it must be special in the sense that the plaintiff’s individual circumstances, or membership of the necessary class, rendered her particularly vulnerable (para 112). If the requisite proximity was established, Tipping J said that it would be necessary to address finally the question of policy, but that should be done when all the facts had been examined (para 130).

Elias CJ and Anderson J preferred a more expansive formulation based on the application of Anns.

In Australia, the High Court held in Modbury Triangle Shopping Centre Pty Ltd v Anzil (2000) 205 CLR 254 that the proprietors of a shopping centre owed no duty of care towards visitors to protect them against the risk of attack in the car park by taking steps to see that it was properly lit. The proprietors were not responsible for controlling the behaviour of the attackers, unlike the prison officers in the Dorset Yacht case who were responsible for controlling the activities of the young offenders in their charge.

In Sullivan v Moody (2001) 207 CLR 562 the High Court held that medical professionals and social workers involved in the investigation of child sex abuse owed no duty of care towards the suspects. The court cited the decision of the House of Lords in Hill in support of the proposition that the conduct of a police investigation involves a variety of decisions on matters of policy and discretion, including decisions as to priorities, and that it is inappropriate to subject those decisions to a common law duty of care.

The Irish courts have consistently followed Hill in holding that the police owe no private law duty of care in respect of their investigatory or prosecutorial functions: Lockwood v Ireland [2010] IEHC 403, LM v Commissioner of An Garda Siochana [2011] IEHC 14 and AG v JK, Minister for Justice Equality & Law Reform [2011] IEHC 65.

In relation to the Convention, Ms Monaghan relied particularly on the decision of the Strasbourg court in Opuz v Turkey (2009) 50 EHRR 695. The applicant and her mother suffered repeated violence from the applicant’s partner, which they reported to the police. He was charged with offences including attempted murder and threatening to kill, but he was released on bail. While awaiting trial he murdered the applicant’s mother. He was released from prison pending an appeal, and the applicant complained that she was given inadequate protection. The court held that there were violations of articles 2 and 3 and gender-based discrimination in violation of article 14 read in conjunction with articles 2 and 3. The court concluded that domestic violence towards women was in practice tolerated by the authorities, and that the remedies relied on by the government in its argument did not function effectively.

The claimants and the interveners also relied on the judgment of Green J in DSD v Commissioner of Police of the Metropolis [2014] EWHC 436 (QB). The claimants were victims of a serial rapist. They succeeded in claims brought against the police under the Human Rights Act and articles 3 and 8 of the Convention. There was no claim at common law. In his judgment Green J carried out a detailed analysis of the Strasbourg jurisprudence regarding the nature and scope of the investigative duty of the police under article 3. The claimants and interveners submitted that his analysis strengthens the case for a common law duty of the scope for which they respectively contend. Green J’s judgment is under appeal.

English law does not as a general rule impose liability on a defendant (D) for injury or damage to the person or property of a claimant (C) caused by the conduct of a third party (T): Smith v Littlewoods Organisation Ltd [1987] AC 241, 270 (a Scottish appeal in which a large number of English and Scottish cases were reviewed). The fundamental reason, as Lord Goff explained, is that the common law does not generally impose liability for pure omissions. It is one thing to require a person who embarks on action which may harm others to exercise care. It is another matter to hold a person liable in damages for failing to prevent harm caused by someone else.

The rule is not absolute. Apart from statutory exceptions, there are two well recognised types of situation in which the common law may impose liability for a careless omission.

The first is where D was in a position of control over T and should have foreseen the likelihood of T causing damage to somebody in close proximity if D failed to take reasonable care in the exercise of that control. Dorset Yacht is the classic example, and in that case Lord Diplock set close limits to the scope of the liability. As Tipping J explained in Couch v Attorney-General, this type of case requires careful analysis of two special relationships, the relationship between D and T and the relationship between D and C. I would not wish to comment on Tipping J’s formulation of the criteria for establishing the necessary special relationship between D and C without further argument. It is unnecessary to do so in this case, since Ms Michael’s murderer was not under the control of the police, and therefore there is no question of liability under this exception.

The second general exception applies where D assumes a positive responsibility to safeguard C under the Hedley Byrne principle, as explained by Lord Goff in Spring v Guardian Assurance Plc. It is not a new principle. It embraces the relationships in which a duty to take positive action typically arises: contract, fiduciary relationships, employer and employee, school and pupil, health professional and patient. The list is not exhaustive. This principle is the basis for the claimants’ main submission, to which I will come (issue 3). There has sometimes been a tendency for courts to use the expression “assumption of responsibility” when in truth the responsibility has been imposed by the court rather than assumed by D. It should not be expanded artificially.

These general principles have been worked out for the most part in cases involving private litigants, but they are equally applicable where D is a public body. Mitchell v Glasgow City Council is a good example. The victim and T were secure tenants of D and were next door neighbours. On a number of occasions T directed abuse and threats to kill at the victim, which he reported to D. D summoned T to a meeting and threatened him with eviction, without informing the victim. Soon afterwards T attacked the victim, causing fatal injuries. The victim’s widow and daughter sued D, alleging negligence in failing to warn him of the meeting with T. The House of Lords held that D was not under a duty to do so, applying the principle in Smith v Littlewoods Organisation Ltd. It rejected the pursuers’ arguments that D’s relationship with its tenant T was analogous to the relationship of D and T in Dorset Yacht or that D assumed a responsibility to protect the victim from T. Mere foreseeability was not enough.

It is true that the categories of negligence are never closed (Heaven v Pender (1883) 11 QBD 503), and it would be open to the court to create a new exception to the general rule about omissions. The development of the law of negligence has been by an incremental process rather than giant steps. The established method of the court involves examining the decided cases to see how far the law has gone and where it has refrained from going. From that analysis it looks to see whether there is an argument by analogy for extending liability to a new situation, or whether an earlier limitation is no longer logically or socially justifiable. In doing so it pays regard to the need for overall coherence. Often there will be a mixture of policy considerations to take into account.

From time to time the courts have looked for some universal formula or yardstick, but the quest has been elusive. And from time to time a court has used an expression in explaining its reasons for reaching a particular decision which has then been squashed and squeezed in other cases where it does not fit so aptly.

Lord Wilberforce’s two-stage formula in Anns appeared at first to usher in a new era of development in the law of negligence, in which prima facie liability at the first stage was drawn very widely but could be negated or cut down by policy considerations at the second stage.

The two-stage formula was stated in terms of general application, but it had particular implications for public authorities, because they have a wide range of duties and responsibilities which would be likely to bring them within the first stage of Lord Wilberforce’s formula.

Doubts about the Anns formula were expressed by the High Court of Australia in Sutherland Shire Council v Heyman (1985) 60 ALR 1 and echoed in subsequent English decisions. In Caparo Plc v Dickman [1990] 2 AC 605 Lord Bridge (with whom Lords Roskill, Ackner and Oliver of Aylmerton agreed) emphasised the inability of any single general principle to provide a practical test which could be applied to every situation to determine whether a duty of care is owed and, if so, what is its scope. He said, at pp 617-618, that there must be not only foreseeability of damage, but there must also exist between the party owing the duty and the party to whom it is owed a relationship characterised by the law as one of “proximity” or “neighbourhood”, and the situation should be one in which the court considers it fair, just and reasonable that the court should impose a duty of a given scope on one party for the benefit of the other. He added that the concepts both of “proximity” and “fairness” were not susceptible of any definition which would make them useful as practical tests, but were little more than labels to attach to features of situations which the law recognised as giving rise to a duty of care. Paradoxically, this passage in Lord Bridge’s speech has sometimes come to be treated as a blueprint for deciding cases, despite the pains which the author took to make clear that it was not intended to be any such thing.

The Anns formula was finally disapproved in Murphy v Brentwood District Council [1991] AC 398. The particular question in that case was whether the owner of a house built with defective foundations was owed a duty of care by the local authority which passed the plans. The House of Lords held that he was not. The property was the plaintiff’s home and it would have cost more than half of its value in good condition to repair the damage caused by the defective foundations. Lord Bridge observed that there might be cogent reasons of social policy for imposing liability on the authority, but that the shoulders of a public authority were only broad enough to bear the loss because they were financed by the public at large, and that it was pre-eminently a matter for the legislature whether these policy reasons should be accepted as sufficient for imposing on the public the burden of providing compensation for the plaintiff’s private loss. Similarly Lord Oliver said that it would not be right for the courts to create new principles in order to fulfil a social need in an area of consumer protection where there was legislation.

Similar considerations underlie decisions in cases not about economic loss: see Stovin v Wise [1996] AC 923 and Gorringe v Calderdale Metropolitan Borough Council [2004] 1 WLR 1057. Both were personal injury cases arising from road accidents.

In Stovin v Wise a highway authority knew that a road junction was dangerous and that the cause of the danger could be removed simply and at little expense. A bank of earth on the corner of the junction obstructed the view of motorists turning right from one road into the other. The highway authority did not own the land but had a statutory power to remove the bank. After there had been a number of accidents it decided to take action. It wrote to the landowner with a proposal to realign the junction but did nothing more and the matter went to sleep until another accident happened. A motorist collided with a motorcyclist whom she had not been able to see until it was too late. The motorist accepted liability to the motorcyclist but claimed a contribution from the highway authority for its negligence. At the trial the judge found the highway authority liable and ordered it to pay a contribution of 30%. On appeal the sole issue was whether the highway authority owed to the injured person a duty of care. The House of Lords by a majority held that it did not.

Lord Hoffmann (with whom Lords Goff and Jauncey of Tullichettle agreed) observed that it is one thing for a public authority to provide a service at the public expense, and quite another to require the public to pay compensation when a failure to provide the service has resulted in a loss. Apart from possible cases involving reliance on a representation by the authority, the same loss would have been suffered if the service had not been provided in the first place, and to require payment of compensation would impose an additional burden on public funds. There would, he said, have to be exceptional grounds for a court to hold that the policy of a statute required compensation to be paid because a power was not exercised.

In Gorringe v Calderdale Metropolitan Borough Council the House of Lords held that the general public law duty of a highway authority under the Road Traffic Act 1988 for the prevention of road accidents did not give rise to a private law duty of care to provide road warnings to alert motorists of hazards. Lord Hoffmann (with whom Lords Scott of Foscote, Rodger and Brown agreed) referred to the fact that in Stovin v Wise the majority left open the possibility that there might somewhere be a statutory power or public duty which generated a common law duty, but he went on to say that he found it difficult to imagine a case in which a common law duty could be founded simply upon the failure (however irrational) to provide some benefit which a public authority has a public law duty to provide (paras 31 to 32). He distinguished that situation from cases where a public authority did acts or entered into relationships or undertook responsibilities giving rise to a duty of care on an orthodox common law foundation (para 38).

In some areas, such as health care and education, public authorities provide services which involve relationships with individual members of the public giving rise to a recognised duty of care no different from that which would be owed by any other entity providing the same service. A hospital and its medical staff owe the same duty to a patient whether they are operating within the national health service or the private sector (Roe v Minister of Health [1954] 2 QB 66). A school and its teaching staff owe the same duty to a pupil whether it is a state maintained school or a private school (Woodland v Swimming Teachers Association [2013] UKSC 66; [2014] AC 537). Educational psychology is a professional service linked to education. An organisation which provides an educational psychology service, and its educational staff, owe the same duty to a pupil whether they are operating in the public or the private sector (X (Minors) v Bedfordshire County Council [1995] AC 633).

Besides the provision of such services, which are not peculiarly governmental in their nature, it is a feature of our system of government that many areas of life are subject to forms of state controlled licensing, regulation, inspection, intervention and assistance aimed at protecting the general public from physical or economic harm caused by the activities of other members of society (or sometimes from natural disasters). Licensing of firearms, regulation of financial services, inspections of restaurants, factories and children’s nurseries, and enforcement of building regulations are random examples. To compile a comprehensive list would be virtually impossible, because the systems designed to protect the public from harm of one kind or another are so extensive.

It does not follow from the setting up of a protective system from public resources that if it fails to achieve its purpose, through organisational defects or fault on the part of an individual, the public at large should bear the additional burden of compensating a victim for harm caused by the actions of a third party for whose behaviour the state is not responsible. To impose such a burden would be contrary to the ordinary principles of the common law.

The refusal of the courts to impose a private law duty on the police to exercise reasonable care to safeguard victims or potential victims of crime, except in cases where there has been a representation and reliance, does not involve giving special treatment to the police. It is consistent with the way in which the common law has been applied to other authorities vested with powers or duties as a matter of public law for the protection of the public. Examples at the highest level include Yuen Kun Yeu v Attorney-General of Hong Kong [1988] AC 175 and Davis v Radcliffe [1990] 1 WLR 821 (no duty of care owed by financial regulators towards investors), Murphy v Brentwood District Council (no duty of care owed to the owner of a house with defective foundations by the local authority which passed the plans), Stovin v Wise and Gorringe v Calderdale Metropolitan Borough Council (no duty of care owed by a highway authority to take action to prevent accidents from known hazards).

The question is therefore not whether the police should have a special immunity, but whether an exception should be made to the ordinary application of common law principles which would cover the facts of the present case.

Ms Monaghan has advanced essentially two arguments in support of the interveners’ liability principle. The first is that the nature and scale of the problem of domestic violence is such that the courts ought to introduce such a principle to provide protection for victims and a spur to the police to respond to the problem more effectively. The second is that the common law should be extended in harmony with the obligations of the police under articles 2 and 3 of the Convention.

I recognise fully that the statistics about the incidence of domestic violence and the facts of individual cases such as the present are shocking. I recognise also that the court has been presented with fresh material on the subject. However, I am not persuaded that they should cause the court to create a new category of duty of care for several reasons.

If the foundation of a duty of care is the public law duty of the police for the preservation of the Queen’s peace, it is hard to see why the duty should be confined to potential victims of a particular kind of breach of the peace. Would a duty of care be owed to a person who reported a credible threat to burn down his house? Would it be owed to a company which reported a credible threat by animal rights extremists to its premises? If not, why not?

It is also hard to see why it should be limited to particular potential victims. If the police fail through lack of care to catch a criminal before he shoots and injures his intended victim and also a bystander (or if he misses his intended target and hits someone else), is it right that one should be entitled to compensation but not the other, when the duty of the police is a general duty for the preservation of the Queen’s peace? Similarly if the intelligence service fails to respond appropriately to intelligence that a terrorist group is intending to bring down an airliner, is it right that the service should be liable to the dependants of the victims on the plane but not the victims on the ground? Such a distinction would be understandable if the duty is founded on a representation to, and reliance by, a particular individual but that is not the basis of the interveners’ liability principle. These questions underline the fact that the duty of the police for the preservation of the peace is owed to members of the public at large, and does not involve the kind of close or special relationship (“proximity” or “neighbourhood”) necessary for the imposition of a private law duty of care.

As to the argument that imposition of the interveners’ liability principle should improve the performance of the police in dealing with cases of actual or threatened domestic violence, the court has no way of judging the likely operational consequences of changing the law of negligence in the way that is proposed. Mr Bowen and Ms Monaghan were critical of statements in Hill and other cases that the imposition of a duty of care would inevitably lead to an unduly defensive attitude by the police. Those criticisms have force. But the court would risk falling into equal error if it were to accept the proposition, on the basis of intuition, that a change in the civil law would lead to a reduction of domestic violence or an improvement in its investigation. Failures in the proper investigation of reports of violence or threatened violence can have disciplinary consequences (as there were in the present case), and it is speculative whether the addition of potential liability at common law would make a practical difference at an individual level to the conduct of police officers and support staff. At an institutional level, it is possible to imagine that it might lead to police forces changing their priorities by applying more resources to reports of violence or threatened violence, but if so, it is hard to see that it would be in the public interest for the determination of police priorities to be affected by the risk of being sued.

The only consequence of which one can be sure is that the imposition of liability on the police to compensate victims of violence on the basis that the police should have prevented it would have potentially significant financial implications. The payment of compensation and the costs of dealing with claims, whether successful or unsuccessful, would have to come either from the police budget, with a corresponding reduction of spending on other services, or from an increased burden on the public or from a combination of the two.

In support of the argument that the court should develop the common law to encompass the duties of the police under the Convention, Mr Bowen and Ms Monaghan submitted that consistency between the common law and the Convention should be encouraged and relied in particular on observations of the Court of Appeal in D v East Berkshire NHS Trust [2003] EWCA Civ 1151, [2004] QB 558, paras 79-85.

There are certainly areas where the Convention has had an influence on the common law. Possibly the most striking example is in the law of confidentiality, which the courts have developed to include a partial law of privacy in response to the requirements of article 8 (Campbell v MGN Ltd [2004] 2 AC 457). But two points should be noted about that. First, the common law had long been regarded as defective. It was heavily criticised by Bingham LJ in Kaye v Robertson [1991] FSR 62, but the Court of Appeal held with regret that only Parliament could cure it. The Human Rights Act 1998 provided the means for reform. In debates on the bill Lord Irvine of Lairg, LC made it clear that in his view the Act would open the way to the courts developing rights of privacy through article 8, and so it did. Secondly, development of the law was necessary to comply with article 8, as interpreted by the Strasbourg court.

The circumstances of the present case are different. The suggested development of the law of negligence is not necessary to comply with articles 2 and 3. On orthodox common law principles I cannot see a legal basis for fashioning a duty of care limited in scope to that of articles 2 and 3, or for gold plating the claimant’s Convention rights by providing compensation on a different basis from the claim under the Human Rights Act 1998. Nor do I see a principled legal basis for introducing a wider duty in negligence than would arise either under orthodox common law principles or under the Convention.

The same argument, that the common law should be developed in harmony with the obligations of public bodies including the police under the Human Rights Act 1998 and articles 2 and 3 of the Convention, was advanced in Smith as a ground for holding that the police owed a duty of care to the deceased after he reported receiving threats. Reliance was similarly placed on the approach of the Court of Appeal in D v East Berkshire NHS Trust (as noted by Lord Phillips MR, who had delivered the judgment of the Court of Appeal in that case). Counsel for Mr Smith relied particularly on the analysis of the effect of the Human Rights Act in D v East Bedfordshire NHS Trust at paras 55 to 87: see the reported argument at [2009] 1 AC 225, 240. The argument by analogy with that case which presently commends itself to Lady Hale is therefore not a new argument, but one which failed to persuade the majority in Smith.

The argument was rejected by the House of Lords for reasons given by Lord Hope (paras 81-82), Lord Phillips (paras 98-99) and most fully by Lord Brown (paras 136-139). Lord Brown did not consider that the possibility of a Human Rights Act claim was a good reason for creating a parallel common law claim, still less for creating a wider duty of care. He observed that Convention claims had different objectives from civil actions, as Lord Bingham pointed out in R (Greenfield) v Secretary of State for the Home Department [2005] 1 WLR 673. Whereas civil actions are designed essentially to compensate claimants for losses, Convention claims are intended to uphold minimum human rights standards and to vindicate those rights. The difference in purpose has led to different time limits and different approaches to damages and causation. Lord Brown recognised that the violation of a fundamental right is a very serious thing, but he saw no sound reason for matching the Convention claim with a common law claim. To do so would in his view neither add to the vindication of the right, nor be likely to deter the police from the action or inaction which risked violating it in the first place.

It is unnecessary for the purposes of this appeal to decide questions about the scope of article 3 and I would not wish to influence the Court of Appeal’s consideration of the judgment in DSD v Commissioner of Police of the Metropolis. It does not alter the essence of the argument which was considered and rejected by the House of Lords in Smith. I am not persuaded that it would be right for the court to depart from that decision, which itself was consistent with a line of previous authorities.

In support of the narrower liability principle proposed by Lord Bingham in Smith, Mr Bowen submitted that limitation of a duty of care to A to cases where A has provided the police with apparently credible evidence that she or he is under a specific and imminent threat to their life or personal safety from a person whose identity and whereabouts are known would satisfy the requirement of closeness or “proximity”. But the majority in Smith rejected Lord Bingham’s formula for reasons which remain cogent. It would be unsatisfactory to draw dividing lines according to whether the threat is reported by A or by someone else (for example, in the present case by the man driven home by Ms Michael’s murderer before he returned and killed her); or whether the threat is credible and imminent or credible but not imminent; or whether the whereabouts of the person making the threat are known or unknown; or whether the threatened violence was to A’s person or property or both. As to the first of those distinctions (whether the threat was reported by A or someone else), Lord Bingham’s own position was ambiguous because his formula confined the duty to a case where the threat was reported by A, but he also disapproved the decision in OLL Ltd v Secretary of State for Transport, in which the concerns about the safety of the children and adults at sea were raised by other people.

More generally, I would reject the narrower liability principle advocated by the claimants for the same reasons as the broader liability principle advocated by the interveners. If it is thought that there should be public compensation for victims of certain types of crime, above that which is provided under the criminal injuries compensation scheme, in cases of pure omission by the police to perform their duty for the prevention of violence, it should be for Parliament to determine whether there should be such a scheme and, if so, what should be its scope as to the types of crime, types of loss and any financial limits. By introducing the Human Rights Act 1998 a cause of action has been created in the limited circumstances where the police have acted in breach of articles 2 and 3 (or article 8). There are good reasons why the positive obligations of the state under those articles are limited. The creation of such a statutory cause of action does not itself provide a sufficient reason for the common law to duplicate or extend it.

So far I have been addressing the appellants’ and the interveners’ arguments. Lord Kerr advances an alternative liability principle which he puts in a broader and a narrower form. He acknowledges (at para 144) that for a duty of care to arise it is necessary to identify a feature (or combination of features) which creates (or create) a sufficient proximity of relationship between the claimant and the defendant. The question “Is there a sufficient proximity of relationship?” is a shorthand way of putting the question posed by Lord Devlin in Hedley Byrne [1964] AC 465 at p 525 “Is the relationship between the parties in this case such that it can be brought within a category giving rise to a special duty?” As Lord Devlin observed, the first step in such an inquiry is to see how far the authorities have gone, for new categories in the law do not spring into existence overnight. In the earlier part of this judgment I have examined how far the authorities presently go and have considered whether there should be a new exception to the general principle about omissions to prevent harm being caused by a third party who is not under the defendant’s control.

Lord Kerr’s broader proposal (at para 144) is that “proximity of relationship” in the present context should comprise these elements: (i) a closeness of association between the claimant and the defendant, which can but need not necessarily arise from information communicated to the defendant; (ii) the information should convey to the defendant that serious harm is likely to befall the intended victim if urgent action is not taken; (iii) the defendant is a person or agency who might reasonably be expected to provide protection in those circumstances; and (iv) he should be able to provide for the intended victim’s protection without unnecessary danger to himself.

Lord Kerr notes that this suggested principle might at first sight appear similar to Lord Bingham’s liability principle, but he observes that his principle, unlike Lord Bingham’s, has the ingredient of proximity built into it as part of what has to be established. This is in my respectful opinion a serious flaw. Whereas Lord Bingham identified the factors which he considered should give rise to duty of care in law, Lord Kerr’s proposition requires it to be established that the relationship has sufficient closeness (proximity) to amount to proximity. In this respect it is circular. It leaves the question of closeness or proximity open ended. It amounts to saying that there is a relationship of proximity if the relationship is sufficiently close for there to be proximity.

Lord Kerr says (at para 163) that the nature of the interaction between the parties is critical to the question whether the necessary proximity exists. He goes on to say (at para 166) that this depends on the facts of the particular case and that for this reason his proposition at para 144 is loosely drawn (or, as I would say, circular). It provides no yardstick for answering the question which it poses.

Lord Kerr says that any narrower test would run the risk of producing anomalous results such as the example which he gives at para 165. In that paragraph he posits the case of a person who through the negligence of the police is given a false impression that an assurance of timeous assistance has been given, on which the person relies. If a person is negligently misled by the police into believing that help is at hand, and acts on what she has negligently been led falsely to believe, she would have a potential claim under the Hedley Byrne principle. Whether that was so in this case is the subject of issue 3. There is, however, nothing anomalous in the Hedley Byrne principle itself or in its limitation. The principle established by Hedley Byrne is that a careless misrepresentation may give rise to a relationship akin to contract under which there is a positive duty to act. Lord Devlin spoke of “an assumption of responsibility in circumstances in which, but for the absence of consideration, there would be a contract” and he said that “wherever there is a relationship equivalent to contract, there is a duty of care” (pp 529-530). To extend the principle to a case in which the core ingredients were absent would be to cut its moorings.

However, Lord Kerr goes on to advance a narrower liability principle (at para 168). His narrower proposition is that whether a relationship of proximity exists should depend on whether sufficient information has been conveyed or is available to the police to alert them to the urgent need to take action which it is within their power to take; the information must be specific; and the threat must be imminent. It is critical, he says, that the police know of an imminent threat to a particular individual, and the duty is personalised to the intended victim.

Lord Kerr’s narrower liability principle closely resembles Lord Bingham’s liability principle, which was rejected by a majority of the House of Lords. It presents most of the problems to which I have referred, such as why a duty should be owed to the intended victim of a drive-by shooting but not to an injured bystander; why the threat should have to be imminent; and why the victim of a threatened arson attack should be owed a duty of protection against consequential personal injury, but not the burning down of his home. Lord Kerr rightly says (at para 181) that the police have been empowered to protect the public from harm. They have indeed a duty to keep the peace and to protect property, which applies to all potential victims of crime. Lord Kerr does not subscribe to the interveners’ liability principle, and I cannot see a proper basis for holding there is a private law duty of care within the terms of Lord Kerr’s narrower alternative.

Mr Bowen submitted that what was said by the Gwent call handler who received Ms Michael’s 999 call was arguably sufficient to give rise to an assumption of responsibility on the Hedley Byrne principle as amplified in Spring v Guardian Assurance Plc. I agree with the Court of Appeal that the argument is not tenable. 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. She told Ms Michael that they would want to call her back and asked her to keep her phone free, but this did not amount to advising or instructing her to remain in her house, as was suggested. Ms Michael’s call was made on her mobile phone. Nor did the call handler’s inquiry whether Ms Michael could lock the house amount to advising or instructing her to remain there. The case is very different from Kent v Griffiths where the call handler gave misleading assurances that an ambulance would be arriving shortly.

Lord Pannick submitted that the majority of the Court of Appeal were wrong to uphold Judge Jarman QC’s decision that the article 2 claim should be allowed to proceed to trial. It is a question of fact whether the Gwent call handler ought to have heard Ms Michael say that her former partner was threatening to return and kill her, and, if she could not hear clearly what Ms Michael was saying because of distractions, whether she should have asked Ms Michael to repeat what she was saying. Lord Pannick argued that even if she should have heard those words, it would not have been enough for a reasonable person to conclude that there was a real and immediate threat to her life. That is again a question of fact. It would be rare for this court to reverse concurrent findings of two lower courts on a question of fact and I do not consider that we should do so in this case. On the contrary, I agree with the majority of the Court of Appeal that the question what the call handler ought to have made of the 999 call in all the circumstances is properly a matter for investigation at a trial. It is not necessary to consider separately the position of the South Wales Police, because Lord Pannick helpfully said that if the cross appeal by Gwent Police failed he would not wish to argue for a different disposal at this stage in the case of the South Wales Police.

I would dismiss the appeal and cross appeal.

Three principal reasons have been given for the conclusion that liability should not attach to the police in this case. The first is that a well-established line of authority dating back to (at least) Hill v Chief Constable of West Yorkshire [1989] AC 53 precluded such liability. The second is grounded on what are said to be general principles of common law. And the third depends on considerations of public policy.

In Hill Lord Keith held that at common law police officers owed the general public a duty to enforce the criminal law but there were “no specific requirements as to the manner in which the obligation is to be discharged” (p 59). On that account an intention to create a duty towards individual members of the public could not be readily inferred. But such a duty could, in appropriate circumstances, arise. It was not enough that police could or should have foreseen that harm to an individual would occur. A further ingredient was required. The nature of that necessary ingredient varied from case to case. In Dorset Yacht Co Ltd v Home Office [1970] AC 1004 the ingredient was the special relationship that existed between, on the one hand, prison officers and the borstal boys who carried out the damage to the boats and, on the other hand, between the prison officers and the owners of the yachts. The prison officers had brought the borstal boys into the locality where the yachts were moored. In that way they had created a potential situation of danger for the owners of the yacht. These circumstances supplied the necessary extra ingredient which allowed a finding of liability to be made. No such features were present in Hill - per Lord Keith at 62C. As he pointed out, the perpetrator in Hill was not in police custody at any material time and the victim was “one of a vast number of the female general public who … was at no special distinctive risk … unlike the owners of [the] yachts” in the Dorset Yacht case.

Lord Keith went on to suggest that there was another reason, grounded in public policy, that an action for damages in negligence should not lie against the police. As Lord Toulson has pointed out, Lord Keith expressed that as a matter of immunity. I will consider the public policy arguments in a later section of this judgment and will mention in passing the dichotomy that has arisen as to whether police should not be held liable for the manner in which they discharge their duties because of an immunity or because an extra ingredient is required beyond foreseeability in order to establish negligence against them. In the meantime, it can be clearly stated that Lord Keith’s formulation of the primary basis on which the plaintiff failed was that an extra ingredient such as was present in Dorset Yacht was missing in Hill.

This extra ingredient has been described as a feature which creates a sufficient proximity of relationship between the claimant and the defendant. What “proximity of relationship” connotes has, perhaps understandably, not been precisely defined. It appears to me that it should consist of these elements: (i) a closeness of association between the claimant and the defendant, which can be created by information communicated to the defendant but need not necessarily come into existence in that way; (ii) the information should convey to the defendant that serious harm is likely to befall the intended victim if urgent action is not taken; (iii) the defendant is a person or agency who might reasonably be expected to provide protection in those circumstances; and (iv) he should be able to provide for the intended victim’s protection without unnecessary danger to himself. This might, at first sight, appear to approximate to the ‘liability principle’ articulated by Lord Bingham in Van Colle v Chief Constable of the Hertfordshire Police; Smith v Chief Constable of Sussex Police [2009] AC 225. For reasons that I will give later, I consider that there is a distinct difference between the two.

This test is criticised on the basis that it is circular. But this is true of any test of proximity and of many other bases of liability, as in, for instance, the test of proportionality - something is disproportionate if it fails to strike a proportionate balance. The notion that any proximity standard inevitably involves an element of circularity is not new. In an article entitled, The vulnerable subject of negligence law Int JLC (2012) 8(3), 337-353, at 338-339, Carl Stychin commented:

There is therefore an inevitably pragmatic dimension (or circularity) involved in the proximity principle but this does not destroy its utility as a standard by which liability is to be judged. In a much cited passage, Deane J in Sutherland Shire Council v Heyman (1985) 60 ALR 1, at 55/6 said this about proximity:

Proximity may in many cases add little to the concept of foreseeability but at root it reflects what Richardson J described in South Pacific Manufacturing Co Ltd v New Zealand Security Consultants & Investigations Ltd [1992] 2 NZLR 282, 306, as “a balancing of the plaintiff’s moral claim to compensation for avoidable harm and the defendant’s moral claim to be protected from an undue burden of legal responsibility” which is exactly what has been the aim of the test for liability which I have proposed. For all, therefore, that the test of proximity may be described as circular, it still has a useful role to play. It is clear, for instance, that it was not present in the Hill case. There was, obviously, no proximity between the police and a member of the public killed by a criminal whose whereabouts were unknown and who, apparently, randomly picked out his victim from the female population.

In Brooks v Commissioner of Police of the Metropolis [2005] UKHL 24, [2005] 1 WLR 1495 Lord Steyn suggested that the principle in Hill’s case should be “reformulated in terms of the absence of a duty of care rather than a blanket immunity” (para 27) but he observed that what he described as “the core principle” in Hill had remained unchallenged for many years (para 30). The “core principle” is that there is, in general, no duty of care owed by police to individual members of the public. Significantly, Lord Steyn had recorded (at para 17) the agreement of counsel that the issues in Brooks should be resolved in the framework of the principles stated in Caparo Industries Plc v Dickman [1990] 2 AC 605 and quoted, apparently with approval, what Lord Bridge had said in that case, at pp 617-618:

Nothing that was said in Brooks, therefore, detracts from the proposition that, provided it is fair, just and reasonable that a duty should arise, police will be liable where they have failed to prevent foreseeable injury to an individual which they could have prevented, and there is a sufficient proximity of relationship between them and the injured person.

Lord Steyn set out a number of policy considerations which, he said, militated against converting the “ethical value” of police dealing respectfully with members of the public into “general legal duties of care on the police towards victims and witnesses” - para 30. But I do not construe this passage as casting doubt on the suggestion that where there was a further ingredient, additional to foreseeability, and sufficient to create a relationship of proximity, liability could arise, provided that it is fair, just and reasonable that it should. By general legal duties I understand Lord Steyn to mean a wide-ranging basis for liability. That is not the primary issue on this appeal, in my opinion. I consider that the question whether there is liability for negligence in this case should rest principally on the claim that its particular circumstances provided the extra ingredient required to create the necessary relationship of proximity between the police and the victim and that it is fair, just and reasonable to find that they are liable to the appellants. I shall discuss those circumstances later in this judgment.

The policy considerations which operated in Hill and Brooks were deployed for theoretically different purposes. In Hill Lord Keith set those out as a justification for an immunity for police against a suit for negligence by an individual member of the public. In Brooks Lord Steyn’s array of public policy arguments was designed to sustain the conclusion that there was no general duty of care owed by police to members of the public. But the policy considerations that have been rehearsed in both cases are relevant in deciding whether, in this particular case, a sufficiently proximate relationship existed between the victim and the police and whether it is fair, just and reasonable that they should be held liable. In the context of the present appeal, therefore, I do not consider it particularly relevant whether the police should not be held liable because actions in negligence against them require to go further than conventional negligence claims or because they are immune from liability by dint of their status.

In Van Colle and Smith Lord Bingham at para 42 said that the most favoured test of a defendant’s liability to a claimant for damage caused by a third party was still that which had been articulated in Caparo. This was described by Lord Bingham in this way: “it must be shown that harm to [the claimant] was a reasonably foreseeable consequence of what [the defendant] did or failed to do, that the relationship of [the claimant and the defendant] was one of sufficient proximity, and that in all the circumstances it is fair, just and reasonable to impose a duty of care on [the defendant] towards [the claimant]”.

Lord Bingham went on, of course, to articulate what he described as “the liability principle”. None of the other members of the Appellate Committee agreed with this as a basis for deciding whether there had been negligence on the part of the police. It is important to note the terms of this principle, however, in order to discuss the current state of the law in relation to liability of a defendant for the acts of a third party. This is how Lord Bingham described the “liability principle”, at para 44:

Expressed in this way, the “liability principle” either omits the proximity of relationship and the fair, just and reasonable components of the Caparo test or treats the relay of the information to the defendant as supplying those ingredients. If it is the first of these, one can see how it does not accord with the “core principle” of Hill, although, interestingly, Lord Bingham believed that his liability principle was not “in any way inconsistent with the ratio” in Hill and Brooks (para 45). If Lord Bingham considered that the provision of information of the nature described supplied the necessary dimension of proximity and the prerequisite that it be fair, just and reasonable to found liability, this raises interesting questions as to how those requirements might be satisfied. Before turning to those questions I must say something about the views of the other members of the House of Lords who disagreed with Lord Bingham.

Lords Hope, Phillips, Carswell and Brown gave various reasons for disagreeing with Lord Bingham’s liability principle. Lord Hope at para 77 suggested that its adoption would lead to uncertainty in its application. He asked who was to judge whether the evidence given to the police was credible and whether the threat was imminent. These were, he said, questions which the police have to deal with “on the spot”. If a judge was to review them it would be on an objective basis and this would lead to defensive policing focused on preventing or, at least, minimising the risk of civil claims. Lord Phillips raised what he considered to be practical difficulties in deciding when the principle would apply. Would it apply, for instance, if the evidence emanated, not from the member of the public under threat, but from some other source; and what if the threat was specific, but not imminent, or imminent but not specific? And why should the principle be restricted to a threat to life or physical safety, but not to a threat to property? Lord Carswell agreed with Lord Hope, whilst observing that he would not disagree with Lord Nicholls who, in Brooks, had said that there might be exceptional cases in which liability might be imposed. Lord Brown considered that it would be difficult to limit the liability principle in the way that Lord Bingham had sought to do; he also thought that defensive policing was “inevitable”; and that the police should be protected from proceedings that would involve a great deal of time, trouble and expense.

It will be seen that the reasons given by the majority in Van Colle and Smith partook, for the most part at least, of policy concerns. None of their speeches addressed directly the question why the relationship between Mr Smith and the police was not one of sufficient proximity. Of course, following Hill, it was still necessary, if proximity was established, to consider whether it was fair, just and reasonable to impose liability on the police. And it appears that it was this latter factor which underlay the dismissal of the appeal by the majority. But it seems to me that the question of whether it is fair, just and reasonable is better considered against the background of whether a sufficiently proximate relationship exists. Put simply, if there is proximity, this is likely to have a bearing on whether it is fair to impose liability. Conversely, if there is not proximity, the issue of fairness etc. is likely to be insignificant. Indeed, it has been suggested that it cannot ever be “fair, just and reasonable” to impose a duty on a defendant with respect to a given claimant if the other stages of the Caparo test are unsatisfied - Peel and Goudkamp, Winfield and Jolowicz on Tort (2014) 19^th ed para 5-036.

I believe that it is necessary to return to the true ratio of Hill and Caparo in order to answer the question whether liability for negligence should be imposed on the police in this case. The core principle of both cases is that liability should not attach to the police unless there is a relationship of proximity and it is fair, just and reasonable to impose it. What is meant by proximity; how can a sufficiently proximate relationship be created; and what circumstances make it fair, just and reasonable for liability to be imposed are all central to the resolution of the issue.

Not only does the answer to the question, “is there a proximate relationship” bear on the matter of what is fair etc., what is ‘fair, just and reasonable’ tends to blend with the concept of ‘proximity’. In the New Zealand case of South Pacific Manufacturing Co Ltd v New Zealand Security Consultants & Collins [1992] 2 NZLR 282, 306 Richardson J held that the proximity assessment will, at root, reflect “a balancing of the claimant’s moral claim to compensation for avoidable harm and the defendant’s moral claim to be protected from an undue burden of legal responsibility”. This sounds remarkably like a weighing of what is fair and just as between the parties. And the authors of Clerk and Lindsell on Torts 21st ed (2014), comment at 8-16 that an assessment of proximity “will inevitably overlap with considerations of justice between the parties”.

As to what is ‘fair, just and reasonable’, Lord Browne-Wilkinson in Barrett v Enfield London Borough Council [2001] 2 AC 550, 559 explained:

This passage clearly contemplates that, in deciding what is “fair, just and reasonable”, courts are called on to make judgments that are informed by what they consider to be preponderant policy considerations. Some assessment has to be made of what a judge considers the public interest to be; what detriment would be caused to that interest if liability were held to exist; and what harm would be done to claimants if they are denied a remedy for the loss that they have suffered. These calculations are not conducted according to fixed principle. They will frequently, if not indeed usually, be made without empirical evidence. For the most part, they will be instinctual reactions to any given set of circumstances.

Similar value judgments are required for decisions on proximity. In Alcock v Chief Constable of South Yorkshire Police [1992] 1 AC 310, 411 Lord Oliver stated that “the concept of ‘proximity’ is an artificial one which depends more upon the court’s perception of what is the reasonable area for the imposition of liability than upon any logical process of analogical deduction”. Again these are value judgments, based essentially on what the court considers to be right for the particular circumstances of the case at the time that the appraisal is being made. It is, I believe, important to be alive to the true nature of these decisions, especially when one comes to consider the precedent value of earlier cases in which such judgments have been made. A decision based on what is considered to be correct legal principle cannot be lightly set aside in subsequent cases where the same legal principle is in play. By contrast, a decision which is not the product of, in the words of Lord Oliver, “any logical process of analogical deduction” holds less sway, particularly if it does not accord with what the subsequent decision-maker considers to be the correct instinctive reaction to contemporaneous standards and conditions. Put bluntly, what one group of judges felt was the correct policy answer in 2009, should not bind another group of judges, even as little as five years later.

In light of the dearth of judicial pronouncement on how to approach, at a level of hypothesis, the question of how a relationship of proximity may be said to exist and in view of Lord Oliver’s statement that it is not susceptible of analogical deduction, one might be tempted to say that “it all depends on the circumstances”. But the law tends to prefer some theoretical rules for the incurring of liability and is wary about accepting that liability can arise unless the “particular circumstances” can be seen to conform to a preconceived set of principles. Having said that, the respondents have accepted, correctly in my view, that, although the rule that the police will not be liable to individual members of the public is a general one, it is not universal. And they have cited a number of examples where liability has been imposed, all of which, apart from the voluntary assumption of responsibility category, depend very much on their own facts.

Whether or not it is necessary to propound a set of principles which can be prayed in aid in order to determine if a particular case constitutes an exception to the general rule, it appears to me incontestable that a proximity of relationship can be created by interaction between parties such as potential victim and police. The nature of that interaction, when it has taken place, is critical to the question whether the necessary degree of proximity exists.

It has been recognised that proximity of relationship can exist where there is a voluntary assumption of responsibility by the police but in cases where this issue has arisen, rules have been applied to strictly restrict its ambit. Relying on those cases (Alexandrou v Oxford [1993] 4 All ER 328; Capital & Counties Plc v Hampshire County Council [1997] QB 1004; and Lord Brown’s observations in Van Colle and Smith at para 135) the respondents argue that unless there was an explicit promise by the police that they would attend immediately and that Ms Michael had expressly relied on this, the conditions for voluntary assumption of responsibility would not be in place. Reference is also made to decisions of the American courts (Noakes v City of Seattle 77 Wash App 694, 700; 895 P 2d 842, 845 (1995) and Perkins v City of Rochester 641 F Supp 2d 168 (2009) which, it is said, confirm the approach that there must be an explicit assurance by the police and express reliance on this by the victim.

One must, I believe, question the logic of this position. Should someone in a vulnerable state, fearing imminent attack, who believes that an assurance of timeous assistance has been made when, through negligence on the part of the police, that impression has been wrongly created, be treated differently from another who has in fact received an explicit assurance of immediate help, if both have relied on what they believed to be a clear promise that police would attend and avert the apprehended danger? The fact that an easily imagined example such as this can demonstrate the anomaly of the current state of the law in relation to voluntary assumption of responsibility indicates that a more expansive (or, at least, a more nuanced) approach is warranted. But it does more than that. It also illustrates the undesirability of creating a set of rules that may at first sight appear reasonable but which bring about incongruous results when applied to cases even slightly different from those in contemplation at the time of their conception.

One is driven therefore to the conclusion that the question whether there is a sufficient relationship of proximity must be primarily dependent on the particular facts of an individual case. It is for this reason that the test which I have suggested at para 144 above is loosely drawn. Any more closely defined test runs the risk of producing anomalous outcomes such as that instanced in the preceding paragraph. Unlike Lord Bingham’s liability principle, however, the ingredient of proximity is not omitted or assumed. It must still be established. And, of course, the question must also be addressed whether there are particular policy reasons militating against the imposition of liability in a specific case.

Proximity in this context means, as I have already said, a closeness of association. In the case of the police it must transcend the ordinary contact that a member of the public has with the police force in general. But the notion that it can only arise where there has been an express assumption of responsibility by unambiguous undertakings on the part of the police and explicit reliance on those by the claimant or victim is not only arbitrary, it fails to reflect the practical realities of life. When someone such as Ms Michael telephones the police she is in a highly vulnerable, agitated and frightened state. Is it to be supposed that there must pass between her and the police representative to whom she speaks a form of words which can be said to amount to an express assumption of responsibility before liability can arise? That the incidence of liability should depend on the happenstance of the telephonist uttering words that can be construed as conveying an unmistakable undertaking that the police will prevent the feared attack is surely unacceptable.

Whether a relationship of proximity can be said to exist should be determined by a close examination of all the circumstances with a view to discovering whether sufficient information has been conveyed to or is otherwise available to the police to alert them to the urgent need to take action which it is within their power to take. That the information be specific and the threat imminent are prerequisites of the proximity relationship. This answers at least some of Lord Phillips’ concerns in Van Colle and Smith. Imprecise information or indefinite timing as to the materialising of any threat cannot be enough to stimulate the police to urgent action and, as I see it, this is an essential dimension of the proximity relationship. In essence that relationship entails the engagement of the police to a response which is out of the ordinary and which is a direct reaction to the plight of the individual under threat. It does not matter if the information is received from a source other than the intended victim. What is critical is that the police know of an imminent threat to a particular individual and that they have the means of preventing that threat and protecting the individual concerned. This is personalised to the intended victim and arises because of the quality of the information which the police have and because they have the capacity to stop the attack.

It is suggested that this formulation is “narrower” than the test set out in para 144 above. I do not believe that it is. The test in para 144 involves the relay of information to the police sufficient to alert them to the need to take urgent action. The information must convey to the police the essential message that serious harm is likely to befall a particular victim. The duty is therefore personalised to that individual. Of necessity, to fulfil this requirement the information must be specific. The imminence of the threat is implicit in the requirement that there is a need for urgent action. But all of this is of minor importance. Of greater moment is the suggestion that this formulation gives rise to problems in the practical application of the test.

I will deal with these supposed difficulties in turn. The first is that the duty as formulated unwarrantably distinguishes between “the intended victim of a drive-by shooting [and] an injured bystander”. I confess to some difficulty in understanding why these categories of person should be assimilated. In the case of an injured bystander the police have no notice of impending harm to that individual on which to act. No circumstances exist in which it might be said that proximity between the bystander and the police has been created. Such an individual is, of course, entitled to the protection that the police owe to members of the public generally but, without more, there could be no duty to protect him from stray bullets any more than there could be a duty on the part of firemen to protect passers-by from dangers caused by a fire which they were tackling.

Any principle for liability of the police in their dealings with individual members of the public should seek to strike a measured and careful balance between the interests of the effective administration of policing and the need to protect vulnerable individuals from serious harm. This will inevitably involve drawing lines which can be portrayed as arbitrary. But the supposed arbitrariness of the operation of the principle in practice should not prevent the law from recognising that liability should attach to glaring omissions where grievous but avoidable consequences ensue. Limiting liability of the police to preventing imminent attack which they are able to thwart may be open to the charge of being arbitrary but it provides a workable basis on which they may properly be held responsible without imposing on them an impossible burden.

Likewise, the restriction of liability to personal injury is defensible on this basis. If it is right that persons such as Ms Michael should be owed a duty of care because of the particular circumstances of her plight, the law should not shirk from recognising that basis of liability simply because it can be posited that there is no logical distinction to be drawn between the need to protect property from the need to protect life. In fact, of course, there is ample reason to distinguish between the two situations. It is entirely right and principled that the law should accord a greater level of importance to the protection of the lives and physical well-being of individuals than it does to their property.

It is true that, unlike the Borstal boys in the Dorset Yacht case, the murderer of Ms Michael was not in police custody nor was he under police control at the time that the telephone call from Ms Michael was received. The murderer was clearly identified, however. Ms Michael was his only intended victim. She had sought the protection of the police from the man whom she feared would attack her again and who proved, in the dreadful event, to be her killer. He, as she told the police, had expressed a specific intention to attack her. The police had been also told that he had already bitten and injured her. It is not in dispute that he had made a specific threat to return to her home to attack her again. And she informed the police that his return was imminent. At this stage in the proceedings it must be assumed that if that information had been acted on promptly, police would have arrived at her home in time to prevent the murderous attack on her. If a proximity of relationship can be created where a victim tells police of a specific, imminent attack on her, it is difficult to imagine what more would be required to create such a relationship than these circumstances. In fact, however, on the appellants’ case, there is more. It is now clear that Ms Michael said to the police operator that her ex-boyfriend had threatened to kill her. The operator claims that her memory is that she did not hear the word, “kill”. At this preliminary stage of the proceedings, the claimants are entitled to assert that the case should be dealt with on the basis that the operator either did hear or should have heard Ms Michael say that the threat had been to kill her. In my opinion, there was clearly a sufficient proximity of relationship.

As Lord Toulson states, English law has not generally imposed liability for the acts of a third party because of the traditional rule that the common law did not normally impose liability for pure omissions. A number of significant exceptions to that traditional rule have been recognised, however, as Lord Toulson has said. In particular, the assumption of a duty to take positive action is one such exception. As he has also pointed out, “assumption of responsibility” is in many instances a misnomer because this is in fact a duty imposed by the court.

In my view, the time has come to recognise the legal duty of the police force to take action to protect a particular individual whose life or safety is, to the knowledge of the police, threatened by someone whose actions the police are able to restrain. I am not convinced that this requires a development of the common law but, if it does, I am sanguine about that prospect. Certainly, I do not believe that rules relating to liability for omissions should inhibit the law’s development to this point.

Tofaris and Steel in their article, Police Liability in negligence for failure to prevent crime: Time to Re-think, (Legal Studies Research Paper Series 39/2014, July 2014) define what they describe as the “omissions principle” in the following way: A is not under a duty to take care to prevent harm occurring to B through a source of danger not created by A unless either (i) A has assumed a responsibility to protect B from that danger, (ii) A has a special level of control over that source of the danger, or (iii) A’s status creates an obligation to protect B from that danger.

In support of this principle, Lord Hoffmann in Stovin v Wise [1996] AC 923, 943 said that “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”. As Tofaris and Steel point out, it is at least questionable that it is particularly valuable to the freedom of a public authority that it should be permitted to negligently fail to assist an identified individual who is at serious risk of physical injury. Whereas it is arguable that a private individual’s freedom has an intrinsic value in its contribution to an autonomous life, the value of the state’s freedom is instrumental and lies in the contribution that it makes to the fulfilment of its proper functions.

The common law has historically required professional persons carrying out a skill to do so with reasonable care and skill. As Tindal CJ put it in Lanphier v Phipos (1838) 8 C & P 475, 479:

In all manner of fields if the professional fails to act with due care and skill, he or she will be liable for any damage caused by their negligence. This is justified on a number of bases; it attributes loss to the person who caused it, it locates compensation in the private rather than the public sector and, arguably, the risk of litigation improves professional standards. The principle holds true even where professionals are acting in response to the acts of third parties. Other emergency services can be liable for their negligence, provided there is sufficient foreseeability and proximity (Kent v Griffiths [2001] QB 36, Capital & Counties Plc v Hampshire County Council [1997] QB 1004). Why should the police be an exception?

It is suggested that the police do not constitute an exception but rather that their exemption from liability is soundly based on the general rule that omissions to act (particularly in relation to actions of a third party) do not give rise to liability. I propose, however, that the cases on which this claim rests can be readily distinguished. In none of those cases was there a proximity of relationship such as exists in the present appeal. In Stovin v Wise, for instance, the failure to improve safety at a road junction affected all who used the particular stretch of road. Likewise in Gorringe v Calderdale Metropolitan Borough Council [2004] 1 WLR 1057. Long-standing or pre-existing dangers stemming from actions of third parties such as in Murphy v Brentwood District Council [1991] AC 398 or the geography of the local area which lay within the public authority’s power to mitigate are of a completely different character from cases where a specific, urgently communicated threat has been imparted to the public agency with the resources and capacity (as well as the public duty) to protect the individual against whom it has been made.

To find that no duty arises on the facts of the present case requires us to squarely confront the consequence of such a finding. If the police force had not negligently downgraded the urgency of Ms Michael’s call, on the facts as they are known at present, it is probable that she would still be alive. While the police are not responsible for the actions of her murderer, if the allegations made against them are established, police played a direct, causative role in her death as a result of their negligence. If they were to be found liable for such negligence, would this be so different from the liability of the doctor of a patient who fails to provide life-saving drugs to prevent an aggressive condition in the necessary time? The police have been empowered to protect the public from harm. They should not be exempted from liability on the general common law ground that members of the public are not required to protect others from third party harm; such protection of autonomy for individuals is not appropriate for members of a force whose duty it is to provide precisely the type of protection from the harm that befell Ms Michael. This is the essential and critical obligation of the police force. Any other professional would be liable for inaction with such grievous consequences. So also should be the police.

Where police have been informed that a member of the public is about to be attacked and they have the capacity to prevent that, the proposition that they should not be held liable because of public policy considerations should be subject to the test which Lord Lowry articulated. Is the anticipated “harm” to the public incontestable? Is it based on tangible grounds rather than mere generalities? Has the burden of establishing the proposition been discharged?

I agree with Lord Toulson that it is difficult to predict with confidence what the operational consequences would be if liability for police negligence was recognised. But the difficulty in predicting whether problems may be encountered should not prompt a refusal to recognise a liability which, by all conventional norms, should be found to exist. A large part of that difficulty stems from the lack of empirical evidence to support any of the feared outcomes such as have been adumbrated in Hill, Brooks and Smith and Van Colle. The lack of empirical evidence led to the Canadian Supreme Court’s distinguishing of Hill in Hill v Hamilton-Wentworth Regional Police Services Board [2007] 3 SCR 129, 2007 SCC 41. And the absence of such evidence was also a key factor in the decision to remove the immunity of advocates in Arthur JS Hall & Co v Simons [2002] AC 615, Lord Steyn at 682D describing the claim that fear of unfounded claims might have a negative effect on the conduct of advocates as “a most flimsy foundation, unsupported by empirical evidence”.

The Law Commission’s Scoping Report on Remedies against Public Bodies (2006) also commented on the lack of empirical evidence to support or contradict the claim that recognition of liability for police negligence would result in a diversion of manpower – paras 3.52-53. At the very least, predictions of a worsening in standards as a result of the availability of judicial review were not borne out. In para 4.25 of its full report Administrative Redress: Public Bodies and The Citizen (2010) (Law Com No 322), the Law Commission referred to the study published by Professor Sunkin and others which suggested that judicial review litigation may act as “a modest driver to improvements in the quality of local government services”.

Set against the poverty - or complete absence - of evidence to support the claims of dire consequences should liability for police negligence be recognised is the fundamental principle that legal wrongs should be remedied. Sir Thomas Bingham MR in X (Minors) v Bedfordshire County Council [1995] 2 AC 633 said that the rule of public policy which has first claim on the loyalty of the law was that wrongs should be remedied. And as Lord Dyson said in Jones v Kaney [2011] 2 AC 398, at para 113:

I do not consider that policy reasons sufficient to displace this general rule have been established. I would therefore allow the appeal.

I would dismiss the cross appeal for the reasons given by Lord Toulson.

In what circumstances can the police owe a duty of care to protect an individual member of the public from harm caused by a third party? There are said to be two objections to imposing such a duty. The first is the “core principle”, recognised in Hill v Chief Constable of West Yorkshire [1989] AC 53 and refined in Brooks v Metropolitan Police Commissioner [2005] UKHL 24 [2005] 1 WLR 1495, that the police are not liable for negligence in the course of investigating or preventing crime. That principle is no longer regarded as an “immunity”, but as a situation in which, for policy reasons, no duty of care is imposed by the law. The second is the general principle in the law of negligence, referred to by S Tofaris and Steel, in their paper on Police Liability in Negligence for Failure to Prevent Crime: Time to Rethink (University of Cambridge Legal Studies Research Paper Series, Paper No 39/2014), as the “omissions principle”, which they describe (I believe accurately) thus, at p 5:

The second objection is the more serious, for there would be little point in considering the strength and validity of the policy reasons which led the House of Lords to formulate the core principle in Hill, and to apply it, not only in Brooks, but also (by a majority) in Smith v Chief Constable of Sussex Police [2008] UKHL 50, [2009] AC 225, if the claim were in any event bound to fail under the ordinary principles of the law of negligence. Those principles, as we see, do not deny any liability for omissions, but impose it only in limited circumstances. Thus, for example, a parent may be liable for failing to feed, clothe, house or otherwise protect her child from harm: see Barrett v Enfield London Borough Council [2001] AC 550. This is because the status of parent imposes a positive duty, probably at common law but certainly under section 1 of the Children and Young Persons Act 1933, to care for one’s children.

But what of public authorities? They certainly owe positive duties towards the public as a whole, or towards certain sections of the public, but do they ever owe a duty of care in negligence towards individuals who suffer harm if they fail to perform those duties? The answer given in cases such as Stovin v Wise [1996] AC 923 and Gorringe v Calderdale Metropolitan Borough Council [2004] UKHL 15, [2004] 1 WLR 1057, is that generally speaking they do not. However, there are exceptions, and one which I find particularly instructive in this case is that established by the Court of Appeal in D v East Berkshire NHS Trust [2003] EWCA Civ 1151, [2004] QB 558.

The House of Lords had held, in X (Minors) v Bedfordshire County Council [1995] 2 AC 633, that a local social services authority owed no duty of care towards an individual child whom it had failed to protect from abuse or neglect in her own home. This was despite the existence of a clear statutory duty to protect such children. However, following the enactment of the Human Rights Act 1998, a child who suffered serious harm as a result of such a failure might well have a claim for breach of her Convention rights: just as the state has a positive duty under article 2 to protect individuals from a real and immediate risk to life of which it knows or ought to know, it also has a positive duty under article 3 to protect individual children from a real and immediate risk of serious ill-treatment or neglect of which it knows or ought to know: Z v United Kingdom 34 EHRR 97, E v United Kingdom (2002) 36 EHRR 519. Thus a court hearing such a claim would have to examine the same factual issues which it would have to examine in a negligence claim: “In these circumstances, the reasons of policy that led the House of Lords to hold that no duty of care towards a child arises … will largely cease to apply. Substantial damages will be available on proof of individual shortcomings…” ([2004] QB 558, para 81).

The court had earlier ([2004] QB 558, para 31) adopted the summary of those policy reasons given by May LJ in S v Gloucestershire County Council [2001] Fam 313, 329-330. These bear a remarkable resemblance to the reasons put forward for the “core principle” in Hill and the later cases. The first, that it would cut across the statutory scheme for child protection, which depended upon multi-disciplinary co-operation, does not apply to policing (and in any event was a dubious reason in child care cases, as the statutory responsibility lay clearly with the local authority). The next four, that the task of child protection is “extraordinarily delicate”, that there was a risk of a more cautious, defensive approach, that it would divert resources away from providing the social services themselves, and that there were other remedies for maladministration, all have their parallels in the police cases. The last, that the development of novel categories of negligence should proceed incrementally by analogy with existing categories, begs the very question at issue.

In the result, therefore, the Court of Appeal held that there was no longer any good reason to deny the existence of a duty of care in negligence towards a child harmed by the failure of a local authority to take appropriate protective action. There was no appeal to the House of Lords against that aspect of the decision (the appeal against the holding that no duty was owed to the parents of a child who was mistakenly taken into care was unsuccessful).

The parallels with this case are striking. There is no doubt that the police owe a positive duty in public law to protect members of the public from harm caused by third parties. In Glasbrook Brothers Ltd v Glamorgan County Council [1925] AC 270, the House of Lords held that the police have a duty to take all steps which appear necessary for keeping the peace, for preventing crime and for protecting from criminal injury. The House also approved a statement by Pickford LJ in Glamorgan Coal Co Ltd v Glamorganshire Standing Joint Committee [1916] 2 KB 206, 229, that a party threatened with violence from another is entitled to protection, whatever the rights and wrong of their dispute. That this is a duty recognised by the common law rather than imposed by statute should if anything strengthen rather than weaken the possibility that it may also give rise to duties towards individuals in negligence.

Equally, there is no doubt that the police may be liable under the Human Rights Act if they fail in their duties under articles 2 or 3 of the European Convention on Human Rights. This part of the claim is to be sent to trial. The issues under the Human Rights Act 1998 are not identical to the issues under the law of negligence, but the existence of a human rights claim means that the policy reasons advanced against the imposition of a duty in negligence claim have also “largely ceased to apply” in a case such as this, where it is alleged that a tragic death would have been averted had the police reacted appropriately to Ms Michael’s emergency call.

It is for those reasons that I would support the analysis put forward by Lord Kerr: the necessary proximity is supplied 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. Once that proximity is established, it is fair, just and reasonable to expect them to take reasonable care to prevent the harm. This is very close to, though somewhat narrower than, the test proposed by Tofaris and Steel (para 189, above). But it is right to acknowledge the strength of the arguments which they so carefully develop, in particular the inter-relationship between the special status and powers of the police to prevent crime and protect people from harm and the limits placed by the law on the ability of people to protect themselves:

However, in developing the law it is wise to proceed on a case by case basis, and the formulation offered by Lord Kerr would be sufficient to enable this claim to go to trial at common law as well as under the Human Rights Act 1998. It is difficult indeed to see how recognising the possibility of such claims could make the task of policing any more difficult than it already is. It might conceivably, however, lead to some much-needed improvements in their response to threats of serious domestic abuse. This continues to be a source of concern to Her Majesty’s Inspectorate of Constabulary: see Everyone’s Business: Improving the Police Response to Domestic Abuse (2014). I very much regret to say that some of the attitudes which have led to the inadequacies revealed in that report may also have crept into the policy considerations discussed in Smith (by Lord Carswell at para 107 and Lord Hope at para 76). If the imposition of liability in negligence can help to counter such attitudes, so much the better. But the principles suggested here should apply to all specific threats of imminent injury to individuals which the police are in a position to prevent, whatever their source.

I would therefore have allowed the appeal as well as dismissing the cross-appeal.