Old Chestnuts, New Problems: A Comment on Tindall v Chief Constable of Thames Valley

An ‘old chestnut’ is an idea or subject that has been discussed for such a long time that it becomes familiar, well-established and unexciting. The opening sentence of the Supreme Court’s judgment in Tindall v Chief Constable of Thames Valley introduces the act/omission distinction in the tort of negligence as having long been recognised, and the court did not see itself as laying down new law.[1] Yet the decision raises new issues for the negligence liability of public authorities for omissions that are far from uninteresting. This case note is structured as follows. Section 1 introduces the legal background to Tindall and the decision itself. Section 2 compares Tindall to another landmark decision in public authorities’ omissions liability, Michael v Chief Constable of South Wales.[2] I argue that Michael would have been decided differently if the interference principle cemented in Tindall had been applied to it. In other words, though the rationes of the decisions are compatible, Tindall provides new grounds for arguing that Michael was wrongly decided. Section 3 considers implications for the future of public authorities’ omissions liability.

1. Tindall v Chief Constable of Thames Valley

1.1 Legal background

English law starts from the position that a person (A) is under no general duty of care to prevent harm from occurring to another through a source of danger that A did not create.[3] From this, three exceptions have been carved out: A does come under such a duty where (i) A 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.[4] Per the equality principle, the requirements of negligence liability for private individuals and public authorities are identical.[5] When public authorities have failed to prevent harm, courtroom battles have been most bitterly fought over whether an assumption of responsibility (exception (i)) can be found.[6] This is because public authorities rarely have a ‘special level of control’ over dangerous third parties (exception (ii)),[7] being a public authority simpliciter does not give rise to special duties of care (exception (iii)), and legislation rarely creates a private cause of action.[8] 

1.2 The decision 

Enter Tindall. Martin Kendall’s car slipped on an imperceptible layer of black ice on the road, causing it to roll into a ditch.[9] Kendall was not seriously hurt, and while waiting for the police, fire and ambulance crews to arrive, he waved down other road users to encourage them to stop or slow down.[10] When the police arrived, he informed them of the black ice, and they put a ‘police slow’ sign on the side of the road.[11] After he was taken away in the ambulance, the police officers left the scene and took the sign with them. No more than half an hour later, with neither Kendall nor a sign to alert drivers to the black ice, Carl Bird lost control of his car in precisely the same area and collided with Malcolm Tindall’s car.[12] Both drivers died on impact. 

The key question was whether the police owed a duty of care to protect Tindall from such harm. The Supreme Court held that they did not.[13] In so doing, the court unanimously recognised the 'interference principle': if (i) A knows or ought to know that B is in need of help to avoid some harm, and (ii) A knows or ought to know that A has done something to put off or prevent someone else helping B, then A will owe B a duty to take reasonable steps to give B the help B needs.[14] This is not an exception to the omissions rule, but an instance in which A’s conduct makes matters worse.[15] Applying this principle to the facts, the Supreme Court held that condition (ii) was not made out—the police did not know and could not reasonably have foreseen that their attendance would displace Kendall’s attempts to warn other road users of the hazard. This is because his intentions to warn other road users were ‘purely private’ and not communicated to the officers.[16] As no exception to the general rule against liability for omissions could be made out,[17] the police did not owe Tindall a duty of care.

2. Old chestnut or new bloom?

In one sense, Tindall is not groundbreaking. The Supreme Court did not see itself as laying down new law, but simply articulating a principle latent in the law since Kent v Griffiths (No 3).[18] However, Tindall could open a side door to public authorities’ omissions liability, offering a way around the ‘assumption of responsibility’ inquiry. This is best illustrated by considering how the ratio of Tindall could have been applied to the facts of Michael.

Joanna Michael called the Gwent Police call centre at 2:29 am to report that her ex-partner had got into her home, violently attacked her, and, before leaving, stated his intention to return shortly to kill her.[19] Michael’s call was wrongly graded on the police’s priority scale.[20] Even though her house was no more than five or six minutes' drive from the nearest police station, 22 minutes elapsed between her call and the police’s arrival.[21] In that time, Michael was stabbed to death in front of her children.[22] The Supreme Court found that the police had not assumed responsibility to Michael as the call handler neither made representations as to how quickly the police would respond, nor advised Michael to remain in her house.[23] The majority’s decision was rightly criticised in Lord Kerr’s dissent and subsequent literature as exceedingly narrow, out of kilter with Hedley Byrne,[24] and holding the assumption of responsibility test ‘hostage to the precise words used’.[25] A way around the limitations set in place by the majority, however, may be furnished by considering the possible application of the interference principle in Tindall to these facts.

Question 1: Did A know or ought A to have known that B was in need of help to avoid some harm?

Plainly so. Joanna Michael reported that her ex-partner had already attacked her, made a specific threat to attack her again, and that his return was imminent.[26]

Question 2: Did A know or ought A to have known that they had done something to put off or prevent someone else helping B?

This requires a factual inquiry that did not take place at trial, but the answer could have been affirmative. Had the police not taken Joanna Michael’s call, she may well have acted differently. She may have called family, friends or neighbours for help. They may have come to her aid. This counterfactual analysis is even more likely in cases where there is a pre-existing tripartite arrangement between the victim, police and third parties. For instance, in Woodcock v Chief Constable of Northamptonshire,[27] the parties had an arrangement that the claimant’s neighbours would alert the police if they spotted her abusive ex-husband.[28] The claimant’s neighbour called the police when he was spotted loitering outside her home precisely because this safety plan was in place. Had the police not entered this arrangement, in a factually similar case, a neighbour might have phoned the claimant instead.[29]

Further, although the interference principle is articulated in Tindall as focusing on whether a third party has been put off from helping B, there is a strong case for extension to instances of claimant reliance.[30] In other words, the interference principle should also catch fact patterns where A’s conduct induces B to refrain from taking self-help measures. Nolan provides the example of a drowning swimmer who, believing the defendant to be on his way to rescue him, swims towards the defendant instead of the shore.[31] There are two reasons to support this view. First, the defendant’s conduct operates identically to make things worse where the counterfactual aid would have been provided by the claimant himself as opposed to a third party. Second, the Supreme Court in Tindall is itself not committed to the alternative help coming from a third party. It cites with approval Capital & Counties plc v Hampshire County Council,[32] in which the defendant fire brigade was found to have made things worse by turning off an automatic sprinkler system, which led to a fire spreading.[33] If the interference principle bites when the help comes from a third party, and when it comes from no party, it cannot sensibly disappear when the help comes from the claimant himself. If this is true, then the case for the principle’s application to Michael is even stronger: had the police not taken Joanna Michael’s call, even if she would not have called upon others for help, she would likely have done everything in her power to flee and/or block entry into her home. 

Tindall demonstrates that a duty of care could have been made out in Michael without engaging the ‘assumption of responsibility’ exception at all. This is interesting for three reasons. First, Kent v Griffiths,[34] which the court in Tindall treats as an early application of the interference principle,[35] was decided before and relied upon by the court in Michael.[36] If the interference principle has been part of English law all along, this offers yet another charge that Michael could and should have been reasoned differently. Second, the court in Tindall relied upon Michael throughout the decision,[37] indeed declining to recognise a new status-based duty of care in light of the ‘weight of that authority and the further body of authority since founded on’ Michael.[38] That the court did not acknowledge the potential inconsistency with the result in Michael, however, is surprising. Third, if above analysis of the interference principle is right, this ostensibly cuts against the policy concerns articulated by the majority in Michael. One reason offered by Lord Toulson for his narrow view of the ‘assumption of responsibility’ exception was that more tort litigation, successful or not, would strain the police budget.[39] Yet Tindall sets a lower hurdle for claimants to clear: it is difficult to imagine cases of reported specific threats where the police ought not to have known that a caller needed protection, and the caller would not have sought alternative help if not for the police answering their call. If floodgates are a legitimate concern, Tindall has cranked the lever.

3. A game of conkers: What future for Tindall and Michael?

We are left with a tension between two Supreme Court decisions: one recently decided unanimously, and one firmly cemented in the tort of negligence. It is thus worth considering how a future court might attempt to resolve this tension if a case attempts to apply Tindall reasoning to facts similar to Michael. 

One possibility is that Tindall and Michael are treated as perfectly compatible as is, simply offering alternative bases of liability. Since Tindall makes it easier to establish police liability, claimants’ counsel can anticipate being met with the full force of the floodgates objections articulated in Michael. For the court, one way around this is to treat the floodgates considerations in Michael as obiter and thus a non-issue. This would make a big leap in the trend of incremental efforts to expand public authorities’ omissions liability,[40] including the recent confirmation that reliance is not always a prerequisite for an assumption of responsibility to bite.[41] However, this would be a surprising departure from the Supreme Court’s express aversion to opening the floodgates when considering other issues in the tort of negligence (most recently and clearly on the matter of psychiatric injury).[42]

Another possibility is to carve out different spheres for Tindall and Michael. One way of distinguishing the cases is that the police’s gross failures in Michael fell within the realm of their core or central functions (responding to emergency calls reporting crime and imminent crime), whereas in Tindall the failure related to their ancillary operations (responding to non-emergency calls; warning road users of hazards). A future court might be of the view that floodgates considerations bite harder in the former category of activity, since the police might carry out such tasks more frequently and the consequences of failure are typically more severe, making litigation more likely to grow unmanageably. However, recognising such a distinction would necessarily undermine the principle of equality, since the core/ancillary activity division is unknown to private individuals’ liability. For instance, there is nothing in the case law to suggest that different rules for Hedley Byrne liability apply depending on whether a company is making representations about core or ancillary terms of service. 

4. Conclusion

While the Supreme Court saw Tindall as merely elucidating a principle already latent in English law, the decision has significant implications for the future of public authorities’ omissions liability. The interference principle, if maintained with full force, provides a viable route for future claimants to establish a duty of care without engaging the difficult and narrow ‘assumption of responsibility’ inquiry. In other news: old dogs learn new tricks, chestnut trees bear new fruit, and the act/omission distinction, established as it may be, continues to produce new tensions in tort law. 

Footnotes:

[1] Tindall v Chief Constable of Thames Valley Police, [2024] UKSC 33, [2024] 3 WLR 822 [1]. 

[2] ibid [56]. 

[3] [2015] UKSC 2, [2015] AC 1732.

[4] ibid [189], citing Tofaris and Steel, ‘Police Liability in Negligence for Failure to Prevent Crime: Time to Rethink’ (University of Cambridge Faculty of Law Research Paper No 39/2014).

[5] ibid.

[6] Robinson v Chief Constable of West Yorkshire Police, [2018] UKSC 4, [2018] AC 736 [32].

[7] See Michael (n 3); HXA v Surrey County Council, [2023] UKSC 52, [2024] 1 WLR 335; Chief Constable of Northamptonshire v Woodcock, [2025] EWCA Civ 13, [2025] 1 WLUK 100 (CA).

[8] Michael (n 3); HXA (n 7).

[9] Gorringe v Calderdale MBC [2004] UKHL 15, [2004] 1 WLR 1057.

[10] Tindall (n 1) [5].

[11] ibid.

[12] ibid [6]–[11].

[13] ibid [12]. 

[14] ibid [88].

[15] ibid [50], [56].

[16] ibid.

[17] ibid [67]. See also academic criticism that the conclusion on foreseeability was hastily reached in absence of a full trial: Eleni Katsampouka, ‘The Omissions Doctrine after Tindall v Chief Constable of Thames Valley Police’ (2025) 88 MLR 825. 

[18] ibid [74]–[87].

[19] ibid [56]: ‘Despite the authors’ recent misgivings, we consider that the “interference principle” articulated by McBride and Bagshaw is a correct statement of English law. Although there has been no previous English case clearly accepting and applying this principle, it is an alternative way of rationalising the result in Kent v Griffiths [2001] QB 36 and one which this court has endorsed in Darnley [2019] AC 831’

[20] Michael (n 3) [5].

[21] ibid [10]–[11].

[22] ibid. 

[23] Jayne Morgan, ‘Joanna Michael: Family's victim support worker speaks’ (BBC, 22 February 2015) < https://www.bbc.co.uk/news/uk-wales-31524938 > accessed 7 November 2025.

[24] ibid [138].

[25] Hedley Byrne v Heller [1963] 3 WLR 101, [1964] AC 465. Hedley Byrne allows for the possibility of implied assumptions of responsibility: see 529, 530; Poole BC v GN [2019] UKSC 25, [2020] AC 780 [80].

[26] ibid [165]–[168]; James Goudkamp, 'A Revolution in Duty of Care?' (2015) 131 LQR 519.

[27] ibid [173].

[28] [2023] EWHC 1062 (KB), [2024] 1 All ER 157 (HC) [108].

[29] ibid [108].

[30] Although this finding was not made on the specific facts of Woodcock: Woodcock (n 7) (CA) [124]. The evidential support for this inference is open to doubt, given that the neighbour’s statement of her unwillingness to ‘get involved’ was made in the context of a call urging the police to intervene.

[31] See Donal Nolan, ‘The Liability of Public Authorities for Failing to Confer Benefits’ (2001) 127 LQR 260.

[32] ibid.

[33] [1997] QB 1004, [1997] 3 WLR 331; Tindall (n 1) [26]–[32].

[34] ibid.

[35] [2001] QB 36, [2000] 2 WLR 1158.

[36] Tindall (n 1) [56].

[37] Michael (n 3) [138].

[38] Tindall (n 1) [75]–[77].

[39] ibid [86].

[40] Michael (n 3) [43].

[41] cf Hill v Chief Constable of West Yorkshire [1989] AC 53, [1988] 2 WLR 1049 63G.

[42] HXA (n 7).

[43] Paul v Wolverhampton NHS Trust [2024] UKSC 1, [2024] 2 WLR 417 [249].