Negligence — duty of care (Caparo and after)

Negligence is the central organising tort of the first-year Tort Law module. It is concerned with careless conduct causing legally recognised harm. The claimant must ordinarily prove duty, breach, factual and legal causation, remoteness, and actionable damage. This week concerns only the first of those elements: whether the defendant owed the claimant a duty of care of the relevant kind.

Duty is not a loose inquiry into whether the defendant behaved badly. It is a legal control device. It identifies the relationships and situations in which the law is prepared to require one person to take reasonable care for another’s interests. In many ordinary accident cases the existence of a duty is now routine: drivers owe duties to road users; employers to employees; doctors to patients; occupiers to lawful visitors. In those cases, the real dispute will usually be breach, causation or quantum. Duty becomes difficult where the claimant asks the court to extend liability: pure economic loss, psychiatric injury, omissions, public authority failures, statements relied upon by third parties, and liability for the acts of others.

The modern starting point remains Caparo Industries plc v Dickman [1990] 2 AC 605. Its familiar three-stage formula asks whether harm was reasonably foreseeable, whether there was sufficient proximity between claimant and defendant, and whether it is fair, just and reasonable to impose a duty. But it is essential, from Week 1, not to treat Caparo as a mechanical checklist capable of solving every negligence problem. The Supreme Court has repeatedly insisted that the law proceeds by established categories and incremental analogy. Caparo is most useful where the case is novel or borderline; it is not a licence to reopen settled duties.

For Durham purposes, this topic is foundational. The first year is compulsory and doctrinally integrated: Contract teaches voluntary obligation, Constitutional Law and Individual and State teach institutional responsibility, and Tort asks when civil responsibility is imposed without agreement. Duty of care is where those themes first meet. It raises technical private-law questions, but also institutional questions about judicial law-making, public expenditure, defensive practices, and the allocation of loss.

A strong answer on duty should do three things. First, identify whether the case falls within an established category. Secondly, if it does not, reason carefully from analogy, not from broad moral intuition. Thirdly, keep duty separate from breach and causation. The fact that harm was foreseeable does not itself establish a duty; the fact that the defendant could have acted differently does not necessarily mean the law required action; and the fact that the claimant suffered loss does not answer whether that loss is recoverable in negligence.

The law of negligence did not begin as a general principle. Nineteenth-century tort law was organised around recognised forms of action and specific relationships: carrier and passenger, doctor and patient, occupier and visitor, manufacturer and immediate purchaser. Liability often depended on privity, possession, or a pre-existing category. The great twentieth-century development was the emergence of negligence as a general, though never unlimited, principle of civil liability.

Donoghue v Stevenson [1932] AC 562 is conventionally treated as the symbolic birth of the modern duty of care. A manufacturer of ginger beer was held capable of owing a duty to an ultimate consumer who had no contract with him. Lord Atkin’s neighbour principle supplied a moral and juridical vocabulary: one must take reasonable care to avoid acts or omissions likely to injure persons so closely and directly affected by one’s conduct that they ought reasonably to be in contemplation. Donoghue did not abolish categories, but it loosened their grip. It allowed negligence to develop beyond contract and beyond rigid status relationships.

After Donoghue, the courts wrestled with the scale of the principle. One danger was excessive restriction: if liability remained trapped in old categories, deserving claimants would be denied recovery merely because the precise relationship was new. The opposite danger was uncontrolled expansion: if foreseeability alone generated duties, negligence might become an open-ended insurance system. The history of duty of care is the history of trying to avoid both extremes.

Anns v Merton London Borough Council [1978] AC 728 represented the high-water mark of generalisation. Lord Wilberforce proposed a two-stage approach: first, ask whether there is a sufficient relationship of proximity such that carelessness may foreseeably cause damage; secondly, consider whether any policy considerations negative or limit the duty. In practice, this was often read as creating a presumption of duty once foreseeability and proximity were shown, subject to policy exceptions. It encouraged expansion, especially in cases involving public authorities and defective buildings.

The retreat came in Murphy v Brentwood District Council [1991] 1 AC 398, where the House of Lords overruled Anns in relation to local authority liability for defective buildings and rejected recovery for the cost of repairing a dangerous defect before it caused personal injury or damage to other property. The loss was treated as pure economic loss. Murphy signalled a broader judicial concern that Anns had made duty too easy to establish and had blurred the distinction between injury to person or property and economic disappointment.

Caparo followed this corrective movement. It did not simply replace Anns with a new universal code. Rather, it reflected a more cautious methodology: foreseeability, proximity, and fairness were relevant considerations, but the law should develop incrementally and by analogy with established authority. Caparo itself concerned auditors’ accounts relied upon by an investor seeking to acquire shares. The House of Lords refused to impose a duty owed to investors at large for such investment decisions. The purpose for which the accounts were prepared mattered.

The post-Caparo period is marked by refinement. In cases such as Marc Rich, Customs and Excise Commissioners v Barclays Bank, Michael, Robinson and Darnley, appellate courts have clarified that Caparo is not a ritual incantation. The first question is whether there is an established principle. Only if the case is genuinely novel does the court undertake the broader evaluative inquiry. This is now the safest way to present the law: duty is category-led, analogy-sensitive, and controlled by policy only where necessary and explicitly justified.

The first principle is that duty of care is a question of law. It is decided by the court, not the jury, and it concerns whether the defendant was under a legal obligation to take reasonable care to avoid the kind of harm suffered by the claimant. It is therefore both relational and scope-specific. One does not owe a duty ‘in the air’. A defendant may owe the claimant a duty to avoid causing physical injury, but not a duty to avoid causing a particular kind of economic loss. Caparo itself illustrates this point: auditors may owe duties for some purposes, to some persons, but not to all foreseeable investors for all investment decisions.

The second principle is foreseeability. Reasonable foreseeability asks whether a reasonable person in the defendant’s position would have foreseen a real risk of the type of harm suffered by the claimant. It is a necessary condition in many duty cases, but rarely sufficient. If foreseeability alone were enough, liability would be vast. A negligent statement in a public document may foreseeably influence many people; a police failure may foreseeably leave a victim exposed; a bank error may foreseeably cause commercial loss. The law still asks whether the relationship is one in which responsibility should be imposed.

The third principle is proximity. Proximity is not mere physical closeness. It is a legal conclusion expressing sufficient closeness between defendant and claimant. It may arise from assumption of responsibility, direct dealings, control, reliance, vulnerability, undertaking, creation of danger, or a recognised status relationship. In negligent misstatement, proximity often depends on the defendant knowing the statement will be communicated to a known person or class for a known purpose and relied upon without independent inquiry. In physical injury cases, proximity is often simpler: motorists and pedestrians are plainly proximate in the relevant sense.

The cases should be read as a sequence, not as isolated slogans. Donoghue v Stevenson supplies the general moral grammar of negligence: a manufacturer may owe a duty to a consumer outside contract where careless manufacture creates a foreseeable risk of physical injury. It establishes that negligence can move beyond privity, but it does not make foreseeability the whole law.

Anns then attempted to generalise duty through a two-stage approach. Its attraction was simplicity: once proximity and foreseeability were present, the burden shifted to policy reasons against liability. Its weakness was that it encouraged courts to recognise duties before fully defining the scope of the loss, the class of claimants, or the institutional implications. That weakness became apparent in defective building and public authority litigation.

Murphy is the decisive correction. The House of Lords treated the cost of repairing a defective building as pure economic loss and rejected the Anns approach. The case restored caution, particularly where the claimant’s loss was financial and where liability would cut across contractual risk allocation or statutory regulatory schemes. Murphy is indispensable because many students wrongly assume that dangerous defects always count as property damage. If the defect has not yet caused injury or damage to other property, the loss is ordinarily economic.

Caparo is the canonical modern authority. Auditors prepared accounts for shareholders as a body, not for investors deciding whether to purchase shares or mount a takeover. The House of Lords refused a duty. The case is less important for its outcome than for its method: foreseeability, proximity and fairness; incremental development; attention to the purpose for which information was supplied. Caparo is strongest in novel duty cases, especially negligent statement and economic loss.

Marc Rich shows Caparo caution in a commercial and regulatory setting. A classification society inspected a vessel. When the ship later sank, cargo owners sued. The House of Lords refused to impose a duty, emphasising the contractual structure, the role of classification societies, and wider policy concerns. The case demonstrates that the law resists imposing duties on peripheral parties where the claimant has other contractual or insurance mechanisms.

Michael concerns police failure to prevent a killing after an emergency call. The Supreme Court refused a common-law duty. The case rests on the general rule against liability for omissions and for failure to prevent third-party wrongdoing, subject to recognised exceptions such as assumption of responsibility. It should not be read as blanket police immunity.

Robinson corrects exactly that misunderstanding. Police officers conducting an arrest knocked an elderly pedestrian to the ground. The Supreme Court held that ordinary negligence principles applied because the officers had positively created a foreseeable risk of injury. Robinson also clarified the status of Caparo: it is not a universal test to be applied in every negligence case, but a tool for novel situations. This is now a central authority for exam answers.

Darnley illustrates established duty in a hospital setting and warns against excessive abstraction. A receptionist gave misleading waiting-time information to an A&E patient who left before treatment and suffered serious harm. The Supreme Court held that the NHS trust owed a duty once the patient was accepted into the A&E system. The case is useful because it shows that administrative staff can fall within the hospital’s duty to take reasonable care not to provide misleading information in the context of patient treatment.

See practice questions section below.

See further reading section below.