THE most troublesome question in negligence today remains omissions liability. The central “exception” is when a defendant assumes responsibility to take positive, protective action – notwithstanding assumption of responsibility's cursory (even dismissive) treatment in the seminal Michael v Chief Constable [2015] UKSC 2, [2015] A.C. 1732. It was a safe prediction that more disputes about its meaning and application would engage the appellate courts. And now, following closely on Tindall v Chief Constable [2022] EWCA Civ 25 (noted Morgan [2022] C.L.J. 245), comes HXA v Surrey County Council [2022] EWCA Civ 1196.

HXA comprised two separate claims by children against local authorities that had failed to take the claimants into protective care when they were being abused by members of their families. The case therefore broadly resembles the leading decision in GN v Poole B.C. [2019] UKSC 25, [2020] A.C. 720. In the High Court, Stacey J. struck out the claims in HXA: [2021] EWHC 2974 (Q.B.). She reminded herself at [64] of the emphasis on precedent and coherence in Robinson v Chief Constable [2018] UKSC 4, [2018] A.C. 736. In this spirit, she held the pleaded claims in HXA to be indistinguishable from GN v Poole, where the Supreme Court clearly laid down that a local authority did not assume responsibility by investigating and monitoring a vulnerable child's situation. The Court of Appeal, however, thought that as allegations of assumption of responsibility “always depend on the specific facts of the case” it would be “plainly wrong” to strike the case out, when the law remained at a “relatively early stage [of] development” since the Poole case: [2022] EWCA Civ 1196 (at [105]–[106]). What emerges?

First, this was an omissions case. Before Stacey J., the claimants made “valiant efforts” to identify negligent acts by the defendant authorities; but the learned judge was dismissive of attempts to “carve out positive acts from a case which is principally about a failure to confer a benefit”: [2021] EWHC 2974 (Q.B.), at [63]. This identified the underlying complaint incorrectly (“failed to see the wood for the trees”). The claimants having been abused not by the defendants but by members of their own families, the complaint was that the defendants could have intervened but did not. This attempt to escape from the Michael jurisprudence was not renewed before the Court of Appeal.

Second, and more difficult, what precisely is “assumption of responsibility”? Many commentators have doubted its coherence, following Barker's scepticism about “Unreliable Assumptions” (K. Barker, “Unreliable Assumptions in the Law of Negligence” (1993) 109 L.Q.R. 461). However, Nolan has strongly argued against sweeping dismissal of the concept: D. Nolan, “Assumption of Responsibility Four Questions” [2019] C.L.P. 123. Whatever the state of the academic debate, in other cases judges have eagerly fallen upon “assumption of responsibility” to circumvent their own restrictive rules on tort liability. In the 1990s it was used to turn the flank of the pure economic loss rule (e.g. White v Jones [1995] 2 A.C. 207; Henderson v Merrett Syndicates [1995] 2 A.C. 145). In the twenty-first century, assumption of responsibility rebuts the “omissions rule” which Stovin v Wise [1996] A.C. 923 revived and Michael (supra) elevated to paramount status.

In HXA the Court of Appeal warned it would be inappropriate to offer general guidance on such a fact-specific question in a striking-out case. They decided merely that on the pleaded facts, assumption of responsibility could arguably be proven. With this important caveat, it nevertheless seems that HXA took a wider view than other leading cases. The court first erased the bright-line rules that Stacey J. had identified in the court below. While it was certainly true that a local authority assumed responsibility to children who had actually been taken into its care (see Barrett v Enfield L.B.C. [2001] 2 A.C. 550), it did not follow that assumption was “confined to cases where it acquires parental responsibility under the Children Act 1989” (at [91]). Where an authority provides accommodation, with parental consent, under section 20 of the 1989 Act, this was not merely an assumption of responsibility about the standard of accommodation, nor was the duty's temporal extent “necessarily confined to the limited period when the child was accommodated” (at [102]).

In one of the cases heard together in HXA the authority had taken a decision to implement measures to protect the children, but then failed to undertake this “keeping safe” work. Stacey J. had been unpersuaded by what counsel presented as “his best point”, holding it “indistinguishable from the reasoning in Poole” [2021] EWHC 2974 (Q.B.) (at [68]). In the Court of Appeal however, Baker L.J. thought it was “at least arguable that, in resolving to take those steps, the local authority was assuming responsibility for the children”: HXA, at [103]. But this seems hard to reconcile with the clear decision in Tindall, supra, that an “ineffectual attempt” to remove a danger does not generate an obligation to continue the attempt (i.e. a positive duty of care). If actually commencing the work, but doing it incompetently and incompletely, creates no duty, why should one arise at the prior stage of resolving to undertake such action? Moreover, in Stovin v Wise (supra) such reasoning was prominent in the dissenting speech of Lord Nicholls but did not dissuade Lord Hoffmann from his famous restatement of the omissions principle. Lord Nicholls emphasised that Norfolk County Council, being well aware of the dangerous road junction, had “decided to act” (to acquire land adjacent to the highway) and had “never changed its mind” – the matter had simply been “allowed to go to sleep” and “overlooked”. This aspect of Stovin v Wise was not discussed in HXA. Tindall was not even cited.

Important factual differences exist. At [101]–[102], Baker L.J. emphasised how common it is for local authorities to be involved with vulnerable children “over a number of years” with the circumstances “varying widely” – thus what a “complex exercise” it is to identify assumptions of responsibility. The facts of the highways cases Stovin and Tindall were simpler, lacking any such extended interaction between claimant and defendant. Another observation concerns the “Diceyan axiom”. Robinson v Chief Constable (supra) proclaimed and implemented the precept, vital for the Rule of Law, that public authorities have no special immunity from negligence claims: when an individual would owe a duty of care (on the facts of Robinson to avoid barging into passers-by in the public street) then so do public officials such as the police. The sting in the tail for claimants is that public officials do not have any additional liabilities either, at common law (exemplified by Michael). That makes sense when police (etc) are doing things that individuals do all the time (driving cars) or might plausibly be doing (chasing fleeing thieves: Robinson), or could just about conceivably be doing (receiving a call for help from a desperate victim of assault: Michael). But the analogy between governmental and private defendants breaks down entirely in family cases like GN v Poole and HXA. It seems most unlikely that any private individual would intrude on another family's affairs over a period of years, as local authority social workers are frequently required to do. And even if, in violation of social mores, this were to happen, a “concerned individual” obviously lacks the extraordinary legal powers enjoyed by local authorities, ultimately permitting removal of children from their parents. We can put it this way. A brave enough private individual might have stepped in to protect Joanna Michael against her former partner's violent attack (pace the observation that the police have a virtual “monopoly” on force: S. Tofaris and S. Steel, “Negligence Liability for Omissions and the Police” [2016] C.L.J. 128); but no citizen could have taken the claimants in HXA into protective custody.

HXA at [98] approved Deakin's observation (noting GN v Poole at [2019] C.L.J. 516) that the fact-specific nature of “assumption of responsibility” means duties of care, in this area, are less abstract questions of law and closer to fact-dependent questions about careless breach. Indeed, many have suggested that a suitably exacting breach threshold would control liability better than the broad-brush of duty of care. In HXA (ibid.) Baker L.J. emphasised that local authorities enjoy a wide discretion in making their complex and difficult decisions – so that there is a “high hurdle” in the way of proving breach. Perhaps this insight should be generalised. It might remove the need to grapple with the mysteries of “assumption of responsibility” and duty of care. The Law Commission thought so in 2008 when it proposed “sufficiently serious breach” as the touchstone of public authority liability: Law Com CP 187, “Administrative Redress” (2008). The Government of the day, eyeing a looming financial crisis, rejected a proposal that might end up damaging the budgets of public services. The climate for reform in the political and economic ruins of 2022 is equally inauspicious. And so a political question – when should child protection services pay for their failings? – remains wrapped within the enigma of assumption of responsibility, or at best deemed a pure question of “fact”.