THERE ARE NEIGHBOURS AND THEN THERE ARE NEIGHBOURS: WHEN SHOULD A JUDGE NOT JUDGE THE POLICE FOR NOT POLICING?
AG v TIO^[*] and TIO v BEENGO et al ^[**]

IAN FRASER^[1]

INTRODUCTION

This is a Comment on a case decided by the High Court and Court of Appeal of Kiribati, in the winter of 2003. It concerns a small jurisdiction, and it occurred some time ago, but the subject was the question posed in the Comment’s title and for several reasons, among them the jurisdiction and the date, this is a subject of some interest, perhaps soon of some lively interest, to anyone involved with the law of the South Pacific – in the courts and in the streets and villages. For it concerns not only the relations between our judges and our police but those among our judges, and between them and their brethren abroad.

THE FLOW OF LAW, AND A CITE UNSEEN

One of the features of the common law that makes it a complex legacy for the South Pacific insular jurisdictions is its sensitivity to local conditions and values. This is a feature not necessarily of its actual responsiveness to society, on which opinion must vary, so much as of its permitted, even necessary, logic, declared by its stewards and authors. In the application of the common law to a dispute it is always a legitimate argument that a potential ruling does not accord with local conditions, or norms and values; and for a great many common-law rules the perceived demand of the local community is the entire justification, the very foundation.
That community is, in the law’s origins, English; and in the law’s ongoing development as received in the Pacific, it is overwhelmingly British and Australasian. British or Australasian, rather; this very sensitivity has generated a series of divurgences in common-law doctrines between Australia and the UK, and even between Australia and New Zealand. As decisions accumulate on a novel issue, or a new extension of the law, or in the adjustment of a traditional doctrine, the jurisdictions are alert to the directions taken within each other, but the path finally taken is the one local courts assert to be most appropriate to local societies.
This process depends on two resources that are scant or even absent in the insular jurisdictions of the Pacific: judges confident of their appreciation of local conditions and values; and litigants in the abundance required by a law that relies on the reiteration and variation of concrete cases. So it cannot proceed as naturally, at least, in the Pacific as in its home societies. Yet Pacific constitutions require that local judge-made law take inspiration (sometimes primary inspiration) from English common law.^[2]

And that allows a manageable role to Pacific courts. In the areas of law where all paths are foreign, courts here maintain fidelity to their constitutional responsibility, in effect, by locating the legitimacy of their rulings in the constituting legislature’s choice of foreign law as a main source of non-statutory law. Within that foreign law, the common law of English-descended societies, legitimacy derives rather from asserted congruence with social norms and values; here, where that law is extended to the islands, legitimacy is a positivistic derivation from the constitutions.
Perhaps the parole rule of contracts suits Melanesian needs as well as it does English needs, perhaps it does not; but in any event Melanesian constitutions direct their courts to take the English path where legislatures have failed to make the law, and so the courts do. The rule, after all, flows naturally enough from contract principle, and beyond doubt the constitution adopts such principles. In torts perhaps the fault principle suits Pacific societies, perhaps it does not, but again it is naturally inherent to the English principles of civil responsibility explicitly adopted by the constitutions. A judge can in good faith assume its appropriateness by relying on legislative supremacy, rather than on the doctrine’s original basis in community standards.
Thus the rules of common law developed in other contexts, by principles originating in other societies, flow with apparent ease through Pacific law, despite the adopted common law’s own nature. You do not need a local precedent to know that a suit in false imprisonment will require proof of total confinement rather than mere obstruction, or that one in negligence claiming a novel duty of care will turn on foreseeability of injury and proximity between the parties.
But this Kiribati case, Tio v Beengo et al and its appeal A-G v Tio, is such a case, a claim to a novel duty of care in negligence. The judges indeed were concerned with foreseeability and proximity, as these terms are used in English precedents. Why need we lead with a disquisition on the jurisprudential basis of Pacific common law?
Because some common-law rules are not based on principle. The principles of the common law are based on social conditions and norms, as the judges have seen these, but some rules are not even said to be based on principle, the principles adopted by the constitutions. They are, rather, said to be the result of policy choices: choices that English or Australian or New Zealand judges have made in explicit isolation from their law’s principles. These rules, instead, are based directly on the particular social context of their application. They are not formalisations of cultural precepts found in local society, but exercises in direct social engineering. For a Pacific court, therefore, legitimacy in applying these rules is not available from the constitutional injunction to base the local law on the principles of English common law.
And the novel duty in negligence claimed by Mr Tio was a duty to take care in protecting him and his interests from criminal acts by third parties, a duty owed to him by the Kiribati police. In English common law, such a claim encounters ‘the Hill immunity’, a rule denying a duty of care in police forces and officers concerning ‘the investigation and suppression of crime’.^[3] It is based explicitly on policy concerns relating to policing in England, a rule self-consciously formulated and applied as an exception to principle. Depending on how a judge construes the facts of Tio’s case, his or her decision must include a decision on whether to apply this ‘immunity’ – and to do that, he or she must proceed from the norms, values, and conditions of Kiribati, just as did the English courts from those of the UK.
Now, there are other such policy-based rules in the common law(s), even others within the tort of negligence. Notably there is the English formula limiting liability for psychological harm suffered by people affected by injury done to third parties by defendants, which is acknowledged even in application to be unsatisfactory yet is maintained in the name of ‘policy’ – and has recently been rejected for Australia by the High Court.^[4] There is, too, another negligence immunity, that of advocates toward their clients, which has recently been abandoned in England and maintained in Australia in decisions turning almost entirely on ‘policy’.^[5] Why focus on the policy rule concerning police immunity?
Because the salience in the South Pacific of psychological-injury negligence claims by ‘secondary victims’ and of malpractice suits against lawyers for their court work is, so far, largely theoretical. Such claims will emerge sometime, no doubt. But claims in negligence against the police, concerning their work in investigating and suppressing crime, have already emerged. Tio’s case does not stand alone. Indeed on an insular scale we could say such claims are proliferating; there have been at least five in this century, in Fiji, Tonga, and Kiribati, all but one to appellate level.^[6]
Moreover this tide flows in Australia too, where (apart from Tame, mentioned above) there were at least three cases in 2004 and 2005 raising ‘the English immunity’ of police operations.^[7] England itself has seen Hill challenged and qualified repeatedly since the decision’s issue, including three appeals to the Law Lords this century.^[8] The doctrine’s status is undecided, indeed contested, in Australia, as common law. In England its status as common law is secure, albeit qualified since Hill, but the implementation of the European Convention on Human Rights leaves it just as contestable as in Australia, as law.^[9]
So Tio’s case participates in an interesting wider set of developments in the common law, including a movement within the South Pacific. But A-G v Tio was decided in August 2003, almost three years ago at the present writing. Since then, just within the region, there have been two other appellate decisions on the police immunity. Why devote a Case Comment to Tio’s case now?
Because it is not really a movement, the set of South Pacific cases dealing with police immunity in negligence. The issue is clearly acquiring prominence, attracting plaintiffs and appeals – but the cases are handled one at a time. They do not form a ‘line’ of cases in the common-law sense, for an awareness of the others is in the observer only: not in the judicial opinions.
The Court of Appeal of Fiji has heard and decided two cases since A-G v Tio in which a police immunity has been asserted and denied by the two sides, the court ruling on the issue in both cases; yet Tio was not cited in either one. But then, in Tio itself, earlier Fiji cases on point were not cited. Nor was an earlier decision on point of the Tongan Court of Appeal. In fact, none of these decisions cited any one of the others: not across jurisdictions, and not even within Fiji.^[10] Case law was considered, indeed relied on, in every judgment, but in every judgment it was case law from the English and Australian jurisdictions.^[11]
So there seems to be a role for academic discussion of this line of cases that isn’t a line – neighbours not of one lain, in the Tok Pisin sense – and Tio’s case is a suitable one to focus on. Its age is a reason to discuss it now, given its status, like the others, of precedent lost. Moreover, it was decided, at trial and on appeal, in ways that expose more clearly than the other cases both the police immunity issue and the deeper problem of whether and how to adopt policy-based common-law rules.

TRIAL: SIMPLY NEIGHBOURS

In November of 1999 two police officers on Butaritai, a Kiribati island, learned that men of the island’s village had resolved to punish a man, Tio, for reasons never advanced in court. They advised Tio to leave his house by 6 pm that day; otherwise ‘anything may happen to you’. When he asked whether they would look after his house if he did, they said they were not sure but would ask their superior, the other police officer on Butaritai. On the way back from his house to their station they told the villagers (assembled for another purpose) not to damage the house, but otherwise, they did nothing about the matter, then or later. That evening, villagers smashed up Tio’s boat and stole his fishing equipment. Tio watched from another house where he had taken refuge.
Tio sued some of the villagers, the three police officers, and the A-G as representative of the police force. (The villagers never appeared, and Tio obtained a default judgment against them.) The officers and the A-G advanced the immunity declared by Hill v Chief Constable of West Yorkshire, with the English progeny usually cited with Hill.^[12] At the trial, reported as Tio v Beengo, Millhouse CJ put the question to himself squarely: ‘Should I, must I, follow in Kiribati Hill?’ ^[13] Of all the Pacific decisions touching on police liability in negligence, this is the only one to pose and answer the question of whether to adopt the Hill immunity.
Hill may be the English common law, he held, but it lacked firm endorsement in the case law of Australia and New Zealand, and in Canada it was explicitly rejected.^[14] Millhouse CJ concluded that he was not constrained by sheer authority. He was free to consider the Hill doctrine on its merits, which is to say on its policy suppositions. And the policy considerations applicable in Britain did not apply in Kiribati.
Why not? His explanation is a little vague, but it does clearly turn on a distinction: ‘English society is sophisticated and complex: Kiribati society is, thankfully, comparatively unsophisticated and simple.’ Kiribati is simply simpler than the UK.
Now, this could be a matter of sheer scale – a jurisdiction like the Western Isles of Scotland in comparison with the British jurisdiction of the House of Lords. Or it could be the nature of crime and crime investigation and suppression in this small land. Millhouse CJ did specify that as a consequence of this distinction, police liability in Kiribati would not produce the ‘unfortunate consequences’ that it would in the larger country, consequences which generated the policy concerns animating the decision in Hill to create the immunity.
This point might be a reference to the likelihood of a ‘flood’ of cases, the product of a population’s sophistication and its size. It does seem reasonable to suppose that negligence floodgates are under more pressure in the UK than in Kiribati.

But the main point seems not to be that, and indeed Hill does not mention the floodgates argument as such. The main point seems rather to be that the kinds of policing decisions negligence suits would put in issue would be easier to evaluate, in Kiribati, without disrupting proper police discretion. Tio’s case, Millhouse CJ holds, is ‘a good example: what the police officers did, did not do, and should have done, are plain’. The English cases are not like this one, he observes. Again he does not say why, but the implication must be that they concern debateable choices and more complex issues of policing philosophy and policy, issues that the House of Lords in Hill regarded as unsuitable for judicial evaluation.
The sort of errors for which the Kiribati police could expect to be sued, that is, would tend to be straightforward, and easy to evaluate for negligence, whereas those for which English forces could expect suits must tend to be complex and subtle. Litigation concerning the simple Kiribati-style carelessness would not involve searching inquiries into the details of past investigations, nor would the prospect of liability foster a defensiveness in the finer points of detective work. Superficial accounts in court of police foul-ups would suffice, the errors being so evident, and any chilling effect – the ‘detrimentally defensive frame of mind’ the Lords were anxious to avoid creating in police officers – would extend only to fostering elementary competence and good faith.
Tio was a ‘good example’. In it the negligence lay in simply failing to do anything after learning that villagers were planning crimes against Tio, in an incident whose relevant facts occurred in the course of a single day. And truly, few accounts of the sophisticated ways English forces have been alleged to be negligent can be related so succinctly.
So the public policy immunity is unnecessary. The case falls to be decided as a straightforward instance of defendants just failing to make any effort to do their plain duty. The duty appears so plain that Millhouse CJ, reasonably, does not pause to record rulings on the usual elements of a duty of care: foreseeability of injury, obviously met on the facts, proximity between the parties, which likewise could seem plain given the direct contact between them concerning the very injury that occurred (and the role of police generally), and whether a duty would be reasonable and fair, qualities apparently subsumed in finding the duty, overall, to be ‘plain’.^[15] It does seem otiose to bother with such a ‘test’, with such ‘factors’, in a situation so clearly covered by Lord Atkin’s commandment in Donoghue v Stevenson [1932] AC 562 not to injure thy neighbour, by action or inaction, a ‘neighbour’ being anyone ‘so closely and directly affected by my act that I ought to have them in my contemplation’.
A simple case, in a simple land. Now, what makes a case simple, generally, is the omission of facts. Millhouse CJ was subjected to that by the parties’ choices, in particular the defendant police force’s choice, not to lead evidence answering the obvious questions.
There was no evidence as to why the police officers stayed away from Tio’s house, or what might have happened if they had attempted to stop the villagers, or how the villagers responded to the officers’ attempt to dissuade them, or what resources were available to the force as a whole. There was no account of what they were doing, or expected to be doing, that evening. There was no evidence about why the villagers were upset with Tio in the first place, or about such incidents occurring elsewhere in Kiribati. The court did not know whether any of the police came from this island. There was not any evidence about what, if anything, police had done subsequently about prosecuting the men who trashed Tio’s property.
We can, therefore, picture the situation as Millhouse CJ did. We can agree that ‘[i]t would be interesting, but not necessary, to know the background...’ Police officers are aware of a serious crime about to be committed, nearby, including the precise location and identities of the victim and the perpetrators, and they ignore it without explanation. Duty is ‘plain’, and in the absence of evidence to show that what looks like dereliction of duty was not truly just that, he held them to the simplest of professional standards (‘Not good policing.’). Breach thus being clear, he awarded damages.
As his last comment, on liability, he remarked that the police conduct here was ‘morally indefensible’; it was a satisfaction to make it legally actionable too.

APPEAL: NEIGHBOURHOOD MUST BE ASSUMED

The Court of Appeal – Hardie Boys, Tomkins, and Pennington JJA, in a single opinion – approached the case very differently, and not so simply. The issue of a duty’s existence should be dealt with first, they held, before considering the Hill immunity. And on these facts a duty is not plain at all.
Indeed what is missing is neighbourhood: ‘what has variously been described as ‘proximity of relationship’ and ‘neighbourhood’’. In a line of English cases before and after Hill where proximity was held to exist between police forces (or prison services) and members of the public (or individual police officers), they found the common element to be the creation of the risk, from whose realisation the plaintiff suffered, by police officers (or prison guards). It is the creation of such a risk by defendant which, in police cases, establishes the special relationship necessary to a duty of care.
So Tio’s case was like Alexandrou,^[16] one of the English cases where proximity was held not to exist (and cited, unlike the others, in Millhouse CJ’s judgment). There the only link between the plaintiff and the police was an automatic burglar alarm, calling them from his store.
Then the CA adds that the Butaritai police never did undertake to look after Tio’s property. Such an ‘assumption of responsibility’, they imply, supported by some of the English cases, could have established proximity between the police and Tio.^[17] Without the assumed responsibility, and in a case where the injury did not result from police fault – although at the outset of the opinion they remark that they make no comment on causation, since it was not discussed at trial – there can be no duty of care. The police conduct here, they agree with Millhouse CJ, was ‘morally indefensible’; but it could not be a breach of a negligence duty to Tio.
And since there is no duty on these general principles, there is no need even to remark upon the immunity from duty created by Hill, or on Millhouse CJ’s ‘robust view’ of its applicability to Kiribati.

PRECEDENT

The basis of the trial decision not having been ruled or even commented on by the appeal decision, Tio’s case is effectively two precedents: one, Millhouse CJ’s, on the Hill immunity, and another, the CA’s reversing him, on proximity in police cases. Each is significant in several ways, like any decision in an area of ‘developing’ law, but each has one feature that seems of particular importance, to Kiribati and to the region.

There goes this neighbourhood: Millhouse CJ’s decision

One can question the soundness of Millhouse CJ’s distinction between simple cases and complex ones as applied so sweepingly to Kiribati, or indeed as applied to the facts of this case. What made the statement of facts so simple was the paucity of fact. In reality, the police must have had reasons for not interfering with the villagers, and the villagers reasons, perhaps related, for their hostility to Tio.
More importantly, this would not necessarily be just a dispute, such as might arise in an English village. Indeed it almost certainly was not anything like that, for English villages participate in a formal system of government with some effective reach into each street, so that any such ganging-up of people against one person, however popular, would have to be in flat contravention of every applicable law. These facts transposed to an English, or Australian, village, would indeed be seen legally as a simple case, the law of government and crimes displacing the sociological complexities to produce a space of simplicity for the operation of the law of negligence.
But the people who trashed Tio’s property, it is clear enough from the facts found, were representing the village. They held no formal office and were exercising no statutory powers – but if they were representing the village, that is a council of ‘old men’ of the village, they were acting as the village government. For Kiribati recognises custom as a source of law, and by Kiribati custom such old men (the unimane, a council made up of an elder from each household) are the authority of the village. Day to day, this is how rural Kiribati is governed, and to the extent it is sanctioned by the Constitution, this authority is lawful. It co-exists with the police stationed on the many islands, but it is the only local political authority.^[18]
Now the unimane enforce their decisions, when necessary, by customary means – which include a form of banishment. This is executed by ordering the offending person or people to evacuate their home by 6 pm, after which some degree of damage or theft of their property (or some degree of personal violence if the offender has not left) constitutes an aspect of the punishment. It seems, of course, likely enough on a balance of probabilities that precisely this is what happened to Tio.^[19]
This is a rather untidy procedure, and the risk of abuse (however ‘abuse’ would be gauged) seems high. But that is not the point here, except insofar as it highlights the uncertainty of evaluating the legality of a given exercise of the procedure. It is indubitably custom, and custom is indubitably a source of law in Kiribati. The legitimate extent of this customary jurisdiction is controversial, as of course must be any particular exercise of it, including this one – for there are, of course, other norms in the Constitution, including a bill of rights in orthodox terms, and Kiribati has not even formalised the authority of the unimane in a statute whose terms could be fairly readily set against the Constitution (as Samoa has done with the unimane’s cultural predecessor, the village fono^[20]). So no-one can say with certainty whether the attack on Tio was a lawful exercise of customary authority, or an abuse of authority otherwise lawful, or indeed legally nothing more than a mob attack. The basic binary distinction of the law, between lawful and unlawful, which grounds the legitimacy of police power, is made fuzzy by the nature of Kiribati’s conceptually fuzzy legal regime. (In part, this is a consequence of the shortage of litigants already remarked.) Apart from the fact that Tio apparently showed no surprise at the news of the impending attack, which proves nothing one way or the other, and the evidence of ‘the people’ in the village interacting as a coherent group with the police officers, there is nothing at all for an observer in our position to go on.

So picture the situation as it probably appeared to the police. Not only did they face the tension in their position of any small force isolated in a rural area, a tension shared to at least some extent by some police posts in ‘sophisticated and complex’ England. What could be described as a criminally-incited, conspiratorial mob, engaged in vandalism and theft, might rather be constitutionally-sanctioned customary authority in action. Or it might be something in between, or something tending one way or the other. The law the police are obliged to enforce allows actions which, with a subtle shift in fact and (customary) law, it might also prohibit.
In other words the case could just as easily be characterised as a very complex, and extraordinarily delicate, exercise of police discretion in law enforcement. Millhouse CJ was not in a position to treat it as such only because the parties chose to leave the matter opaque – which in itself suggests a degree of sensitivity and subtlety in the officers’ task. It might well have been precisely the kind of decision difficult to explain, let alone portray with proof, that Hill is meant to protect from illegitimate judicial second-guessing: decisions which, in Lord Keith’s terms, ‘would not be regarded by the courts as appropriate to be called into question’.
Indeed the sort of judicial examination a full-blown negligence case would entail might be even less appropriate in Kiribati than in any English incident, given the cultural gap between the court and the islands. Even rendering the situation as perceived by the parties into English would pose serious issues (compare the term unimane, compelling an Anglophone reader’s attention to the exoticism of its referent, with its English rendition, always used by I-Kiribati: the old men). A negligence trial, and a negligence verdict and subsequent orders, could risk a significant chilling effect on the exercise of discretion in policing. For an analogy picture, perhaps, the judicial review in England, for purposes of a negligence claim by one of the parties, of a rabbi’s resolution of a dispute about the governance of an Orthodox synagogue, originally conducted in Hebrew: although, really, that would be far closer to common-law culture than village-level Micronesian governance and its relations with local police officers.
So treating such a case as simple could, and in some cases must, amount to riding one form of Kiribati law – the formal introduced criminal law – roughshod over another form of Kiribati law – the vaguely constitutional indigenous law of governmental authority. It would, to put it mildly, change the neighbourhood. And that would risk the very inversion of another of Lord Atkin’s commandments in Donoghue: that rulings on duties of care must be based on ‘a general public sentiment of wrongdoing’.
Millhouse CJ is not lightly to be taken as oblivious to all this. The point of his judgment seems to be that evaluating that public sentiment in Kiribati, and resolving the abrasive conflicts between customary authority and written criminal law, should be done by the Kiribati legislature. In the meantime, if the police are content to present their case bereft of real context, that is how the court will judge it.

Good fences make good neighbours: the CA’s decision

The effect of the CA’s decision is to leave Tio’s complaint, as a matter of Kiribati common law, within the village. His choice-of-law conundrum is resolved: on facts like these, the common law offers him no cause of action against the police. The involvement of the formal state in the situation does not show in the common law. Apart from the possibility of a legislated remedy, perhaps under the Constitution, he is left to respond to the unimane – by collecting on his default judgment or otherwise.
This could be the point of the judgment too; or it might be just an incident of an adoption of common-law rules, here concerning not the immunity but the tests for a duty of care in negligence. By these tests, the court holds, proximity is established either by the creation by the police of the relevant risk or by their ‘voluntary assumption of responsibility’. Manifestly the police here did not instigate the vandalism and theft that injured Tio’s interests, and, in the CA’s view, their knowledge of the impending attack, their propinquity, and their contact with Tio did not constitute such an assumption.
But these are not clearly plausible holdings. To create a risk must mean, in law, to contribute to causing the risk to exist, and there was no evidence as to the effect of the police inaction upon the villagers’ resolve to attack Tio’s property. Even construed as omission – failure to advertise readiness to interfere and failure to interfere – the role of the police could be taken as causative; and their contact with the villagers, and conveying of the threat to Tio, could equally well be construed as positive action.
Moreover, the precedents cited do not all support the main proposition. In one, Costello v CC Northumbria [1999] 1 All ER 550 (CA), the defendant police officer’s role in the injurious incident was, precisely, merely standing by; that incident was an attack on the plaintiff officer by a third party resisting the plaintiff, an incident which the defendant had done nothing at all to bring about. More pointedly, there is a precedent conspicuously absent from the CA’s selection: Osman.^[21] It, too, was a case of police inaction in the face of a risk they played no positive part in creating. And the English CA’s holding in it was that proximity did exist, due to the police officers’ extensive contacts with the plaintiffs and their knowledge of the particular threat posed by the third party who eventually shot the plaintiffs.
Osman is also a case illustrating the role of assumption of responsibility, fairly clear on its facts, in contrast to another Hill progeny, Alexandrou. The CA in AG v Tio held that Alexandrou was the case among the English precedents which the Kiribati case resembled. An automatic burglar alarm sounding at the police station – the link between the parties in Alexandrou – is thus like the Tio officers advising Tio that the villagers would be attacking his house that night. This is already not intuitively obvious. But in Alexandrou the police had responded, and checked the alarmed premises; because they missed the signs of forced entry (carelessly, Alexandrou alleged) they failed to prevent the subsequent burglary of which the plaintiff complained. That is, they did not know the crime would occur; indeed they had reason to believe, subjectively, that there was no particular likelihood of it occurring. The Kiribati police were at the polar opposite of states of mind. The English CA held that a mere alarm did not establish proximity. If knowledge of the relevant risk is material to proximity, as of course it must be to ‘assumption of responsibility’, then the cases are contraries rather than analogous.
Nonetheless, Alexandrou did declare that the same result would obtain where a member of the public called the police, and this obiter dictum, if not the actual case, could apply analogously to Tio. It depends on whether calling the police is sufficiently like the police calling on you.
The tenuousness of these holdings emerges in the light of the CA’s remark that the police conduct was ‘morally indefensible.’ Now, the police actions in Alexandrou, held to be a case like Tio’s, could hardly be characterised as that. They were careless, rather, as in most of the other cases cited by the Court of Appeal where there was proximity: Home Office v Dorset Yacht Co. [1970] AC 1004 (HL)(HO guards’ sleep allowing Borstal boys to escape via plaintiffs’ yacht), Rigby v CC Northamptonshire [1985] 2 All ER 985 (QB)(failure to have a fire truck standing by when teargas was used to drive an armed madman out of the plaintiff’s shop), Alcock v CC Yorkshire [1992] AC 310 (crowd-control error causing deaths of plaintiffs’ relatives in stadium stampede), Knightley v Johns [1982] 1 All ER 851 (order to drive wrong way along one-way road to block access to street).
Where one can see moral wrong is in the cases of not only proximity but an ‘assumption of responsibility’ – assuring a family that the madman harassing and threatening them will be arrested soon enough while failing to arrest when evidence sufficed and it was clear that a physical attack was likely (Osman), failing to assist a fellow-officer being beaten by a prisoner (Costello, above), and, perhaps, leaving the personal file of an informer loose on the seat of an empty police car (Swinney v CC Northumbria [1996] 3 All ER 449 (CA)).

Indeed in this ‘practical world’ where the Good Samaritan cannot reasonably be the standard,^[2] it is difficult to see how inaction could be called morally indefensible in the absence of at least ‘proximity’, if not the particular form of proximity called ‘voluntary assumption of responsibility’. Carelessness per se is measured objectively, and so has no moral significance; it is the duty it violates, and particularly the relationship between the parties (which is what ‘proximity’ refers to, however vaguely) helping to create that duty, which lends the lack of care a moral significance.
That the police in Tio did not create the risk, if that is a reasonable view of the facts, is not the barrier to proximity that the CA calls it, at least on the English common law they rely on. And their moral judgment belies their legal conclusion, given the fragility of the (only) analogy they drew. A satisfactory interpretation of the decision must then return to its consequences, mentioned above. It maintains a fence between the formal introduced law, here being summoned through negligence doctrine, and customary political authority.

Over Hill with Tio in the Pacific

However narrow its basis in fact, Millhouse CJ’s approach is quite sound as a qualification of Hill. Indeed the qualification is inherent in the Hill doctrine, for it was precisely Millhouse’s distinction that Lord Keith invoked as a ground for the immunity on the Hill facts.
The sort of police decision and action whose ‘policy and discretion’ would be inappropriate for review by a court was contrasted, by Keith, to ‘allegations of a simple and straightforward type of failure – for example that a police officer tripped and fell while pursuing a burglar’. The airing of police conduct, and its subjection to judicial inspection, concerning that sort of allegation, he implies, would pose no issues of public policy.
It is the difficulty of distinguishing such cases from the inappropriate ones, and the exposure of matters inappropriate to litigation risked even by the drawing of the distinction, which justifies the blanket approach adopted by Hill. (The allegations in Hill itself, Keith held, were of the inappropriately complex sort.)
That issue is avoided if a whole jurisdiction’s policing activity may be characterised as overwhelmingly of the simple sort. Where police detective and forensic work just does not include complex investigations and bureaucratic procedures, and consists largely of following simple rules to halt clear crime, apprehend known suspects, and preserve obvious evidence, the prospect and experience of negligence litigation should have no more undue chilling effect than it does on auto mechanics.
So the viability of a ‘simple and straightforward’ exception to the Hill immunity depends on the viability of such a description of a jurisdiction’s policing. In Kiribati and similar jurisdictions that depends on one’s view of what policing should be, for if the task is simply the enforcement of the written law the work is indeed straightforward. Crimes against the person and against tangible property are indeed clear, as a rule, typically involving known individuals and obvious evidence.^[23]
Now, one can identify other kinds of crime in such jurisdictions, and ways in which policing could be more than enforcement of a criminal code whenever the police learn of violations. Is the relative obscurity and difficulty of financial crime weighty enough, and the possibly inscrutable informality of how some matters are left to custom resolution legitimate enough, to reject the bifurcation of partial immunity and partial exposure to liability, as the House of Lords rejected it? Is the distinction between the simple tasks and the subtle ones difficult enough to draw, or the drawing of it so likely to trench upon the matters properly left free of judicial interference, that any doctrine relying on such a distinction should be avoided, as the Lords held in Hill?
These are issues to be settled in the ‘Pacific Way’; we must hope they are not settled by too easy an acceptance of the law of distant neighbours.

^[*] [2003] KICA 10.
^[**] [2003] KIHC 89.

^[1] BA, LLB, BCL (McGill), LLM (Dalhousie); Lecturer, School of Law, University of the South Pacific, Vanuatu.
^[2] Usually directly; sometimes indirectly. In the case of Kiribati, the Constitution is silent as to the law’s other sources; but the Laws of Kiribati Act 1989 (No 10 of 1989) directs the courts first to customary law, then to the common law of Kiribati (ss. 4(2) and 6(3)), defining customary law as ‘customs and usages, existing from time to time, of the natives of Kiribati’ (s.5(1)) and the common law as ‘the rules comprised in the common law, including the doctrines of equity, of England (in this section referred to as “the inherited rules”), as applied in the circumstances pertaining, from time to time in Kiribati’ (s.6(1)). The English borrowing is subject to the usual exclusion of rules that are ‘inapplicable or inappropriate’ to Kiribati (s. 6(2)(b)).
^[3] Hill v Chief Constable of West Yorkshire [1989] AC 53 (HL), the ‘Yorkshire Ripper’ case, holding that in addition to lack of proximity between a police force and a class of potential victims of an unknown serial killer where the class is ‘young women in Yorkshire’, a public-policy immunity protects police forces generally from a negligence suit ‘of this kind’ concerning their actions in ‘the investigation and suppression of crime’.
^[4] Alcock v Chief Constable of South Yorkshire [1992] 1 AC 310 (HL); the Australian rejection is in Tame v NSW (2002) 191 ALR 447 (HC).
^[5] Arthur Hall v Simons [2000] 3 All ER 673 (HL) and D’Orta-Ekenaike v Victoria Legal Aid [2005] HCA 12 (HC) http://www.austlii.edu.au.

^[6] The decisions are, in sequence of issue: Wartaj Seafood Products v Min. Home Affairs & A-G [2000] FJHC 100, Jagroop v Sokai & Tonga [2001] TOCA 10, Bachu & Wati v Comm. Police & A-G [2004] FJCA 53, and Kumar v Comm. Police, Comm. Prisons, & A-G [2005] FJCA 35 (all citations http://paclii.org.vu/).
These are the case which could make up a line of cases: see Fraser, ‘Police Negligence in the Pacific: Building a Case around Hill (Case Comment on Kumar v Comm. Prisons, Comm. Police, & A-G Fiji [2005] FJCA 35’, forthcoming, Commonwealth Law Bulletin.
The clearest indications among them that Hill is adopted are in the lone trial-level decision, Wartaj, and in Kumar (where, however, it is offset confusingly, perhaps confusedly, by conflicting dicta). These indications are obiter.
A few others would lie outside the line. A Fiji plaintiff took a long shot some years ago, in Bokoci (Takayawa Estate) v Kumar et al (Fiji Boxing Council) & A-G [1995] FJHC 143 (http://paclii.org.vu/), vainly alleging a police duty to look after the contestants in licensed boxing matches. And there are Papua New Guinea cases against the police, by members of the public and based on negligence, which do not mention any immunity. But the facts these cases treat as ‘negligence’ are also batteries, indeed homicides, far from the concerns of Hill or Tio.
^[7] In the same year as Kumar in Fiji the NSW Court of Appeal, in Thompson & Thompson v Vincent et al [2005] NSWCA 219, denied the claims but indicated bases on which they might accept one: where there was assumption of responsibility ‘in a particular task’.
In Batchelor v State of Tasmania [2005] TASSC 11 the Tasmanian Supreme Court allowed an action to proceed on the basis that any immunity might not apply where the impugned police act was contrary to the force’s own policies and instructions. And the court in Batchelor leaves open the issue of whether Hill even applies in Australia.
Earlier the NSWCA, in Cran v State of NSW [2004] NSWCA 92, had characterised the argument from Hill as ‘the English immunity’, and suggested (obiter) a police liability where the plaintiff could show an assumption of responsibility – even where the plaintiff is an accused person. (All citations http://www.austlii.edu.au/).
^[8] In Waters v Comm. Police of the Met [2000] UKHL 50 the Lords denied a strike-out application by a force sued by one of its officers for failure to deal with bullying and intimidation by fellow officers. Responsibilities as an employer displaced the immunity in matters not strictly amounting to ‘the investigation and suppression of crime’. Lord Steyn for the majority acknowledged that this was a development – an “increment” to the law – and that it entailed adding to the consumption of police time and resources. Lord Hutton, in a concurring opinion, simply held the Hill policy concerns to be outweighed by the policy-based need to respond to such shocking mismanagement of the Force.
In A-G v Hartwell [2004] UKPC 12 the Privy Council allowed a negligence action against the police of the British Virgin Islands by a member of the public. Mr Hartwell was shot in a bar by a police officer firing wildly at his ex-girlfriend, using the police revolver assigned to the island where he was on duty. The Lords approved this suit not on the basis of vicarious liability, but of direct negligence by the force, in giving the officer unsupervised access to the firearm. This was said to be a ‘positive act’, creating a risk, as opposed to the failure to perform to a standard that is protected by the Hill doctrine.
In Brooks v Comm. Police Met [2005] UKHL 24, the House of Lords approved the Waters exceptions to the immunity – and took care to dissociate itself from ‘the full breadth’ of the Hill doctrine.
(All citations http://www.bailii.org/).
^[9] The Lords in Brooks noted that the advent of these rights’ incorporation into English law weakened the basis of Hill. On 10 March this year, in Van Colle & Van Colle v CC Hertfordshire Police [2006] EWHC 360 (QB) http://www.bailii.org/ a trial judge effectively overruled the Hill immunity, allowing what appears to be the first direct challenge to it by a plaintiff relying on the Convention rights rather than the tort of negligence.

^[10] The absence of cross-citation is not the product of literal ignorance, for the appellate benches of these jurisdictions share judges: Tomkins JA sat on the Tio and Jagroop panels (in Kiribati and Tonga), Pennington JA on Tio and Bachu (in Kiribati and Fiji).
^[11] As well as an occasional insert from Canada or New Zealand. The HL and PC decisions mentioned above at n.7, however, and the Australian ones at n.6, are not mentioned in the Pacific cases.

^[12] A cluster of Court of Appeal decisions spawned by Hill reported in a single All England Report a few years later. An automatic burglar alarm’s notification of the police of trouble at the plaintiff’s store was not significantly different from any emergency call to the police, and these made up far too numerous and disparate a category of relations to constitute ‘proximity’: Alexandrou v Oxford [1993] 4 All ER 328. The sighting by police officers of a potential driving hazard on the road did not put them into ‘proximity’ with all other drivers who came upon the hazard later: Ancell v McDermott [1993] 4 All ER 355.
In each case the availability of the immunity was noted. Only in the third case, Osman v Ferguson [1993] 4 All ER 344, did the facts compel the court actually to invoke it. There, as in Hill, the injury was to the plaintiff was the criminal act of a third party, a man known by the police to be a criminal at large – but this time the police knew who he was, were practically and legally capable of arresting him before the injury, knew that the plaintiff’s family were his particular targets, and had, in addition to all that, explicitly advised the family to rely on their efforts to deal with him. Proximity had to be found, and was, so in Osman the policy immunity was the ratio decidendi for dismissing the action. Osman has been ever since the ultimate example of the immunity’s potency, given the sympathetic appeal of its facts.
The Lords refused leave to appeal Osman, but the Osmans took their case to the European Court of Human Rights, alleging a violation of their European Convention on Human Rights right to access to justice: although the device of striking-out a claim is necessarily implied by the common-law concept of duty as a prerequisite of liability in negligence, it does mean the plaintiff’s case is never heard on the merits, and the ECHR decided this was a denial of judicial process: Osman v UK (1998) 5 BHRC 293. More to the present point, it also decided that the Convention right to life imposed a positive obligation upon police to assist people like the Osmans – the holding inspiring Van Colle, above n.8.

^[13] The opinion, like the CA’s on appeal, is brief enough to omit paragraph citations.

^[14] Beckstead v. Ottawa (City) Chief of Police (1997) CanLII 1583 http://www.canlii.org/, (1997) 37 OR (3d) 62. The Canadian tort of ‘negligent police investigation’ was reconsidered, and approved, in Hill v Hamilton-Wentworth Regional Police et al (2005) CanLII 34230 (Ontario CA), which is now on appeal to the Supreme Court: (2006) CanLII 13738 (SCC).

^[15] So runs the test for novel duties of care established by Caparo v Dickman [1990] 1 All ER 568 (HL). It remains, in a rough way, the test in England, although it has been rejected for Australia (Sullivan v Moody (2001) 75 ALJR 1570 (HC)).
^[16] Above n.11.

^[17] More precisely, the ‘assumption of responsibility’ creates a relationship between the parties, the plaintiff relying on the defendant fulfilling the role it has ‘assumed’ (i.e. implicitly accepted), which one can describe as ‘proximity’ – and it displaces any judicial consideration of whether the duty of care would be fair and reasonable, since, after all, the defendant has already, ‘voluntarily’, accepted it. This last gloss was made explicitly by Williams & Reed v Natural Life & Mistlin [1998] 2 All ER 577 (HL).

^[18] See, for example, Jane’s Oceania Page (Dame Jane Resture) at http://www.janeresture.com. She mentions:

....new social institutions also depend on their success on the village or island authorities to ensure the usefulness in the community. In practice, the determination of their identification and the embodiment with the community is the prerogative of the unimane who will accord the appropriate sanctions. The implementation of decisions by the latter rests finally with the younger generations... (http://www.janeresture.com/ki33/social_structure.htm)

^[19] The ‘six-o’clock order’: personal communication, I-Kiribati student, USP School of Law (April 2006).

^[20] For a case in Samoa similar to Tio’s, in which the village council, or fono, resists liability, see Leituala v Mauga [2004] WSSC 9 http://paclii.org.vu/. The statute is the Village Fono Act 1990 (Samoa No 3, 1990).

^[21] Above n.11.[]

²² The law’s world, as characterised in Lord Atkin’s discussion of the norm of loving one’s neighbour, in Donoghue v Stevenson, leading to the passage about those neighbours being the people one is affecting closely and directly – answering ‘the lawyer’s question’ differently from Luke 10:30-37, yet not quite so harshly as does the ‘assumption of responsibility’ doctrine.
^[23] This statement could be qualified by noting that any dispute or violent incident in the Pacific involving land title is not clear, as a legal issue.
In Thompson & Thompson, above n.8, the defendant police were caught up in a complex issue of land rights (about which plaintiff and another were breaching the peace), and the NSW Court of Appeal took pains to emphasise that police have no obligation to sort out ‘torts exam questions’ as part of keeping the peace. They were reinforcing a distinction between criminal law and land law that in the Pacific becomes willy-nilly one between introduced law and customary law, land being generally held by customary title.