The distinction between acts and omissions runs deep in the common law of torts. At one stage in danger of being overlooked or elided, it is today more prominent than ever before. Defendants are not usually liable for failing to assist, protect or otherwise make claimants better off. Under Michael v Chief Constable of South Wales [2015] A.C. 1732, as explained by Robinson v Chief Constable of West Yorkshire [2018] A.C. 736, this usually entails that public authorities will not be liable in negligence. The sordid business of weighing competing policy concerns is unnecessary. The omissions principle, when it applies, is sufficient to deny liability.
But it is neither universal nor unqualified. First, of course, is the question whether a situation properly concerns an “omission” rather than positive misconduct. If so, liability may still exceptionally arise from the defendant's control over the situation (Dorset Yacht Co. Ltd. v Home Office [1970] A.C. 1004); or, a much further-reaching exception, when the defendant assumed a positive responsibility towards the claimant.
These principles have again been considered in Tindall v Chief Constable of Thames Valley [2022] EWCA Civ 25, [2022] R.T.R. 16. A driver, one Kendall, skidded on a patch of black ice. While awaiting assistance, Mr. Kendall warned approaching traffic about the ice. The police arrived and put up a warning sign. But once accident debris had been cleared from the road, and Mr. Kendall had gone to hospital in an ambulance, the police left the scene, removing the warning sign and making no attempt to alert drivers to the black ice. Shortly afterwards, another car skidded on it, crashing into Mr. Tindall's vehicle. Mr. Tindall was killed. His estate claimed in negligence against the police. It was alleged that the police had created or increased the danger posed by the ice by removing both Mr. Kendall and their own warning sign; and/or had assumed responsibility to road users by their taking control of the first accident. Master McCloud held that a duty of care arguably arose. The Court of Appeal, disagreeing, held that both suggested bases of liability were unarguable and struck the claim out.
Stuart-Smith L.J. reviewed the authorities in a judgment of exemplary clarity (e.g. summary at [54]). In essence, such “ineffectual interventions” are still omissions. They do not make the claimant's situation worse; but merely fail to improve it. Nor can intervention per se constitute “assumption of responsibility”. These points are exemplified by Capital and Counties plc v Hampshire CC [1997] Q.B. 1004. That decision, denying liability solely because of the omissions principle rather than policy reasons that were not “convincing” (see pp. 1042–44), seems surprisingly modern. Unlike numerous House of Lords cases, Capital and Counties survives the Supreme Court's “corrective re-emphasis of the principled orthodoxy” on omissions (Tindall, at [60]). Police cases also illustrate the point. No liability when the police fail to intervene (Michael itself; or Ancell v McDermott [1993] 4 All E.R. 355 (failure to warn about/remove diesel spill on highway, despite knowledge of the hazard)). It is otherwise where the police cause harm or create danger (Robinson itself; or Rigby v Chief Constable of Northamptonshire [1985] 1 W.L.R. 1242 (CS gas canister deployed into building started a fire)).
Not every case is straightforward to categorise. Stuart-Smith L.J. noted at [44] how GN v Poole BC [2020] A.C. 780 “lie[s] somewhere between” Michael and Robinson, “because it was the positive act of the Council in placing the children in the particular property that exposed them to the risk of abuse from the anti-social neighbour, though the claimants also complained of a failure to move them out of [that house]”. Nor does Poole stand alone. In Mitchell v Glasgow CC [2009] A.C. 874, the law lords disagreed whether the council landlords were purely passive bystanders in the dispute between the pursuer and his homicidal neighbour, or had actively intervened by informing the neighbour that he would be disciplined (triggering the fatal assault). Delicate borderline classification mattered less when (as in Mitchell) judges could nevertheless arrive at the same conclusions for public policy reasons. Following the “corrective re-emphasis”, the act/omission distinction becomes dispositive. No wonder that claimants highlight positive conduct within the totality of the defendant's behaviour. However, this tactic may fail: “The attempt to carve out positive acts from a case which is principally about a failure to confer a benefit is to fail to identify correctly the underlying complaint” (HXA v Surrey CC [2021] EWHC 2974 (Q.B.), at [63]). Courts may deplore “semantic bickering” about acts and omissions (Rushbond Plc v JS Design Partnership LLP [2021] EWCA Civ 1889, at [53]), but fine distinctions are unavoidable.
On the “control” exception, Stuart-Smith L.J. stressed (Tindall, at [72]) how far removed the facts were from the “classic” example of Dorset Yacht. It was “far too wide” to suggest that whenever a public authority has power to prevent harm, a duty arises to do so. That would contradict “firmly established principles” and the leading cases of East Suffolk Rivers Catchment Board v Kent [1941] A.C. 74, Stovin v Wise [1996] A.C. 923 and Gorringe v Calderdale MBC [2004] 1 W.L.R. 1057. Quite so. Indeed, one wonders whether Dorset Yacht itself might require “corrective re-interpretation” given the dubious reliance on Donoghue v Stevenson [1932] A.C. 362 (e.g. [1970] A.C. 1004, 1027), violating the precept that one cannot get a duty out of a power. Perhaps Dorset Yacht survives less as a “classic” case, more an outlier (warders’ total control of prisoners); or (as Stuart-Smith L.J. notes) as an “introduction of danger” case, since the Home Office built the non-impermeable prison next to the benighted yacht club in the first place.
To infer an assumption of responsibility from what the police did in Tindall would similarly contradict the governing authorities (see [65], [73]–[74]). On the current state of the case law failure seems inevitable. It would nullify the “ineffectual assistance” doctrine (see above) if every failed intervention amounted to “assuming” an obligation to succeed. Given that there was no interaction whatever between Mr. Tindall and the police, the “assumption of responsibility” case was even more tenuous here.
But it will continue as the focus of attempts to establish positive duties of care. Assumption of responsibility's meaning and application “have not yet been fully settled”, but cases can usefully be divided between general relationships (e.g. between solicitors or bankers and their clients) and “particular relationships created ad hoc” (Hughes v Turning Point Scotland [2019] CSOH 42, at [83]–[84]). One difficult question is why positive duties readily inhere in some “general” relationships but not others (see Michael, at [179] (Lord Kerr); Robinson, at [115] (Lord Hughes)). Crudely put, doctors owe such positive duties but the police do not. The former even extend to a hospital receptionist's duty to provide patients with information about waiting times: Darnley v Croydon NHS Trust [2019] A.C. 831. Since the claimant in Darnley arguably bore some responsibility for leaving the hospital before receiving medical attention (compare [2018] Q.B. 783, at [56] and [2019] A.C. 831, at [29]), liability there contrasts grotesquely with Michael where the claimant – utterly blameless and utterly dependent on police protection against a murderous assault – received nothing. Can the omissions principle bear the weight of such startling differences?
Perhaps after all, “policy” may explain the categories where assumption of responsibility is, or is not, established. Capital and Counties suggests (see p. 1036) that because fire brigades act for the benefit of the public generally they do not assume responsibility to particular property-owners, especially as the public interest and individuals’ interests could conflict (e.g. property sacrificed to form a fire-break). (See also PJSC Uralkali v Rowley [2020] EWHC 3442 (Ch), at [379]–[387]: administrator's duties to insolvent company precluded “assumption of responsibility” to prospective purchasers of the company.) The “conflicting duties” argument is a particular powerful instance of the wider policy concern that liability could incentivise undesirable behaviour (see Hanna Wilberg (2010) 126 L.Q.R. 420). Before the “corrective re-emphasis”, courts routinely relied on the potential distortion of crime-fighting priorities to conclude that the police do not owe positive duties of care (e.g. Van Colle v Chief Constable of Hertfordshire [2009] 1 A.C. 225). That reasoning seems more satisfying than a bare assertion that answering a 999 call or attending the scene “cannot” be an assumption of responsibility (an odd sort of impossibility when it can arise in analogous cases – compare also Phelps v Hillingdon LBC [2001] 2 A.C. 619 (educational psychologists owed duties of care qua professionals) with GN v Poole (social workers owed no duties of care – qua unprofessionals?)).
Not everyone is convinced by Robinson's “corrective re-emphasis” (compare Lord Mance's and Lord Hughes’ judgments in that case). True, policy reasoning has often been cursorily undertaken, using lordly conjecture rather than hard facts. Even at best, policy leads towards pragmatic compromise rather than the straight road of shining principle. But concepts are not ends in themselves. Courts need to ask when and why “assumption of responsibility” should be recognised. Precedent stabilises the common law. But excessive reliance on previous decisions at the highest judicial level is dereliction. The elegant simplicity of the omissions principle should not mask difficult questions about the proper extent of public authority liability.
