Allocation of Risk and Tort Law

Allocation of Risk and Tort Law

8.1 Introduction

In this chapter, we turn our enquiry from contract to tort. Our concern here is with the role of insurance in determining the existence and ambit of tort duties.¹ The chapter extends the concerns of Chapter 7, but the terrain is more diverse, and in terms of received opinion we face a much bigger challenge. That is because, quite simply, influential accounts have denied that insurance plays any role at all in determining the existence and ambit of tort duties.² Indeed, some of the same accounts suggest that it makes no sense to say that tort duties have such a thing as ‘ambit’ at all, except in special cases where they are restrained on ‘ex post’ grounds.³ In conventional cases, they are simply duties to act in particular ways. We will explain our disagreement with these views in the course of the chapter. However, such is the entrenched acceptance of the doctrinal irrelevance of insurance to tort duties, we should be clear from the start that there are certain assertions about insurance which are the target of these exclusionary views, and which are not part of the analysis which we offer. We are, in other words, hoping to move the debate to new terrain, where positions need not be so sweepingly stated and assumptions may be questioned.

In particular, the received view that insurance concerns as a whole are and should be irrelevant to determination of principles of tort liability may be based on a fear that consideration of insurance will lead either to a simple ‘deep pockets’ argument (thus to circularity, as insurance follows liability, which follows insurance), or to tort being swallowed up in its entirety by loss-spreading concerns. Both of these tend to associate insurance concerns with an ex post search for deep pockets. This is not our account of the role of insurance. In fact, we suggest that the role of insurance is complex and context-dependent, and not reducible to one sweeping proposition. Most particularly, it is not to be reduced to an ex post search for a compensation fund, in which duty becomes irrelevant.

The received view (that insurance is doctrinally irrelevant to tort) retreats from the fear that tort will be reduced to a search for deep pockets by suggesting that insurance is essentially a private matter between assured parties, and their insurers. Insurance is best seen as something parties can choose to obtain, or must obtain in some contexts, in order to deal with the liabilities, and is irrelevant to tort questions. An olive branch is proffered, in that it is suggested that insurance may have a role in supporting tort in achieving its goals;⁴ but the end result is that this merely cements its doctrinal irrelevance.⁵ In our view, just as insurance does not offer a single knock-down source of rational answers to tort questions, neither should it be excluded from the doctrinal picture. A starting point to our departure from the received view is that ex ante bargaining around risks is not outside the frame of tort reasoning, and that from this point of view, insurance is by no means to be written off as ‘fortuitous’ as far as other parties are concerned.⁶ Given the centrality of insurance to contractual arrangements, charted in Chapter 7, it is simply impossible to ignore insurance while taking account of party risk arrangements. The extent to which the courts do take account of such arrangements has been understated, whether claims arise as between contracting parties, or on the part of others exposed to risk by the parties’ arrangements. We will further suggest that the risk-allocation analysis of tort duties is more variable and extends further into the law of tort than is generally appreciated, as too does the role of insurance.

Given the entrenched hostility to including insurance considerations in the determination of tort duties, we should underline that we do not consider insurance to be an overriding factor which ‘explains’ the entire law of tort, nor even those parts of the law of tort which are most closely associated with making good material losses. Nor do we consider that the law of tort (or significant parts of it) should be seen as collapsing into one large loss-spreading device operated by the courts: the job of the courts is not to seek the ‘best loss-spreader’ in each case, and they are not equipped to do so.⁷ Nor is our analysis likely to show that tort is a poor relation of insurance in the distribution of losses, for it shows that in practice the relationship is symbiotic. Rather, our enterprise is focused on understanding the internal operation of the law of obligations, and in relation to tort as in relation to contract we find that in many instances risk allocation, and thus (as we have explained in the previous chapter) inevitably insurance, is inherent to this operation.

We would point out that to consider the role of risk allocation in this way is not to deny that tort liabilities are dependent on duties; nor that these duties are imposed by law and reflect a variety of normative concerns. On the contrary, we suggest that those cases where duty is a key issue in tort litigation have frequently raised issues about risk allocation, and about the relationship between tort and other forms of ordering of risk. Through the remedies for harm occasioned by their breach, tort duties have the effect—and are approached by parties, courts, and legislators as having the effect—of allocating various risks to those who bear them. Courts treat this as a reason to be circumspect in the extension of duties.⁸ The fact that the duties are not simply duties to ‘insure’ against loss, but set various standards of behaviour, by no means undermines this point.⁹ In fact, the rich variety of tort duties indicates that risk allocation is a complex matter. Where the duty is to take reasonable care, the risk allocated is of carelessly caused harm, to the extent that it is attributable to the breach and falls within the scope of the duty. So, for example, where the law seeks an ‘assumption of responsibility’ as the prerequisite of a duty of care, the required assumption may (depending on the parties’ relationship) be interpreted as an assurance that the claimant need not worry about the risk of the defendant’s negligence—not just that the defendant will take care.¹⁰ Even in the case of duties which approach the absolute in their content,¹¹ issues arise surrounding the ambit of the duty and whether the particular instance of harm falls within, or outside, the risk thus allocated to duty-bearers; and whether the defendant is a person owing such a duty to the claimant.¹² The existence of different duties throughout the law of tort is not inconsistent with the idea that tort law allocates risk. Duties are a key determinant of the risks to be allocated to various parties, and are often quite deliberately tailored to this effect.¹³

8.1.1 Trajectory: risk structures from contract to tort

In Chapter 7, we argued that courts applying contractual principles attach particular significance to the parties’ allocation of risks, and that insurance and the transfer of risk have been under-emphasized features in these allocations. We argued that predictions of the demise of party risk allocation as a core concern in the law of contract had been overstated and indeed were misconceived. However, our analysis of party risk structures departed from the traditional bipartite frame of classical contract theory. In particular, our analysis underlined the involvement of multiple parties which is inherent where risks involved in contracting are insured,¹⁴ and we emphasized the role of the courts in supporting party arrangements—particularly in respect of the allocation of risks—in their market context. Linked to our analysis in Chapter 5, which indicated the extent of issues surrounding subrogation and the need which often arises for courts to consider insurers in relation to other indemnifiers, this will include evaluative choices about the appropriateness of holding an insurer, or another contracting party, to be the ‘primary indemnifier’ in light of planned risk structures, an issue which is submerged from view because the very existence of an insurer may remain at least superficially concealed, and because traditional bipartite analysis encourages this state of affairs.

We saw in Chapter 7 that predictable liability risks are typically allocated by parties, with varying degrees of sophistication depending upon the nature of their contract. This is significant, in that parties often anticipate and deal with risks caused by fault whether their own, or that of their employees, or of others. It makes sense, in other words, to discuss the risk of fault as one of the risks to be allocated. Certainly, parties do not focus uniquely on deterrence when determining how to deal with such risks. Equally, the law is perfectly ready to support parties when they negotiate around the financial burden of their liabilities and structure their relationships accordingly. Thus, our general portrayal of the role played by the law of contract has affinity with the paradigm of ‘dynamic market individualism’ outlined by Roger Brownsword as typical of the courts’ approach to commercial contracts,¹⁵ in which courts interpret and support party agreements. We saw that not all accounts accept that legal principles are designed to support party risk arrangements. The very idea of a mutual risk structure across a number of parties will remain suspect when viewed within the strongly bipartite frame associated with classical contract doctrine (or with some approaches to the law of tort). Beyond this, it is in any event much less widely recognized that the location of insurance is a key factor in these risk allocations; and a number of cases prove on examination to be brought by insurers whose claim for indemnity would, if it succeeded, upset a planned risk allocation. We argued that the involvement of insurance in the parties’ structuring of risks and duties has been underestimated in theoretical accounts of such cases. But where do tort duties fit into this analysis?

Tort duties are generally conceived of as creations of law and legal policy, rather than as springing from party agreement. As we saw in 7.1, this provided the essence of a ‘welfarist’ argument, developed particularly during the 1970s, that tort was gaining dominance over contract: given the particular approach to tort adopted by its proponents, this was effectively also an argument that legal policy was dominant over party intention.¹⁶ Courts have, however, been vigilant in restraining the potential impact of tort in unravelling contractual allocations of risk. A recent expression of the basic difference can be found in the judgment of Jackson LJ in Robinson v PE Jones (Contractors) Ltd:

Contractual obligations spring from the consent of the parties and the common law principle that contracts should be enforced. Tortious duties are imposed by law, as a matter of policy, in specific situations.¹⁷

It will be noticed that Jackson LJ refers here to ‘specific situations’. We suggest that this well reflects the current position of the courts, and indicates that it is misleading to suppose that tort duties are also different from contract duties in being owed to ‘persons generally’.¹⁸ A first key question, therefore, is whether the existence and definition of such duties are capable of being affected by party arrangements. This can be easily answered, because in at least some instances, it is perfectly clear that they are.¹⁹ It is widely recognized (though with differing degrees of enthusiasm) that party arrangements, particularly party risk allocations, have influenced decisions about the existence and nature of tort duties where the parties ‘come together’ against the background of a contractual structure. Such cases have been loosely captured in the idea that the duties fall to be considered within a ‘contractual matrix’, a term explained in 8.3. Broadly, ‘contractual matrix’ cases are those where contractual structuring of risks forms the background to the proposed tort duty.

The expression ‘contractual matrix’ is generally used outside those cases where the claimant and defendant are directly related by contract (‘concurrency’ cases). Claimant and defendant may, instead, be related by a series of contracts, whether these are best referred to in terms of a ‘chain’, or (borrowing Hugh Collins’ description of construction contracts), in terms of a ‘hub and spokes’.²⁰ In either case, arrangements are in place which structure the location of risks among more than two parties; and the argument is that tort duties fall to be considered against the background of such arrangements. It might be added, however, that the role of the claimant in such cases may vary widely, in that the claimant might be a participant in a common risk structure, or outside that structure and adversely affected by it. The response of the law is rightly sensitive to these differences; and in this respect, sensitivity to party arrangements is not incompatible with protection of consumers or others.²¹ Reasoning associated with the ‘contractual matrix’ may even be relevant where the defendant is not a contracting party, and the question arises of how far such issues extend into the law of tort.²²

That being the case, our second key question is: what role is played in the determination of such duties by insurance arrangements? This question has not had the emphasis it deserves in existing accounts. The analysis in Chapter 7 has shown how deeply insurance concerns are implicated in party allocations of risk; and we draw further on this analysis in 8.3 and 8.4. Continuing the themes of Chapter 7, but extending them to analysis of tort duties, we show that insurance arrangements present different issues from the mere presence, or even possibility, of insurance; and that these are generally part and parcel of the prevailing risk structure, even if there is silence on such issues on the part of most scholars.

It is widely recognized that tort duties are often proposed in such circumstances because the law of contract is hampered by its concept of privity from responding appropriately to multi-party arrangements. The question is whether tort is able to be sensitive to the notion of a planned, multi-party transaction, where the law of contract falls short. It is suggested that tort will hinder, rather than help, in this context if it too is overly fixated on two parties rather than the existence of a multi-party context; if it proves insensitive to existing risk structures; and if its goals are inflexibly defined. Simply shifting from a classical, bipartite analysis of contract, to an equally doctrinaire, bipartite reading of tort, will only succeed in producing a hierarchy of rules or values in which either tort or contract trumps the other, preferred risk structures being the victim. Hence, the turn to tort law to overcome contract’s limitations where third parties were concerned has also necessitated some flexible reorientation of the reasoning to be deployed in relation to tort duties.²³ That reorientation has plainly occurred, even if there is continued scholarly resistance to it. It is not surprising that this has had an effect in other areas of the law.

In relation to ‘contractual matrix’ cases, our goal is to highlight the continued importance of insurance when the focus shifts to tort. This will itself be resisted by some.²⁴ Over a series of articles, Jane Stapleton has consistently rejected any role for insurance in determining whether tort duties are owed in all categories of case,²⁵ sometimes by reference solely to previous work of her own.²⁶ Such is the influence of this body of work that we must engage with it here. As one part of the general project, Stapleton has proposed that a claimant’s alternative opportunity for protection may be a valid reason for refusing a tort duty, but only if that opportunity was consistent with enhanced deterrence of negligent conduct.²⁷ Highlighting (by contrast) our point in the preceding paragraph, this approach rests on a stark division between the purposes of tort and contract duties, while elevating a single proposed purpose of tort duties (deterrence) over that associated with contract (which is more often recognized as structuring of risk).²⁸

Rejecting these arguments, we begin by seeking to redress the neglect of insurance in ‘contractual matrix’ cases involving tort duties. But we also raise some questions about the general depiction of tort law involved in Stapleton’s critique, and contemplate the influence of risk allocation, including insurance issues, beyond a small subset of tort duties. It is salutary to be reminded that the bulk of tort litigation is conducted away from the appellate courts—away from the courts at all. But we would point out that what Stapleton describes as core instances of ‘traditional’ tort claims—road traffic accidents between strangers²⁹—are themselves deeply shaped by insurance and by distributive concerns. They present a very poor paradigm of tort law if what is desired is to expunge the influence of insurance. Both ends of the spectrum—cases where parties have structured risks, and cases between strangers—therefore raise insurance issues, albeit in different ways, and we do not find the existing prohibition of discussion of insurance to be helpful at any point along it.

Reaching further from the domain of contractual structures, we ask whether the existence and shape of specific tort duties is influenced by a much more general recognition that the imposition of tort duties, just like the interpretation of contractual duties, itself has the effect of allocating risk. Even if the source of tort duties is recognized to be conceptually different from the source of contract duties, the impact of such duties is in important respects the same.³⁰ Both tort and contract duties oblige parties to behave in particular ways; but both also, inescapably, allocate the risks associated with a failure to do so. This is the very purpose of many contractual terms. That it is also an inescapable result of tort follows from the fact that the consequence of recognizing a tort duty is that liability will follow as of right where relevant harm is caused by its breach. With a particular focus on the English courts, we suggest that there is increasing emphasis on ex ante allocation of risks, as part of the process of determining the existence and the ambit of tort duties; and we explore the relevance of insurance in this process.

8.1.2 Summary

Shifting our attention from contract duties to tort duties considerably broadens and diversifies the range of our enquiry. The relationships which give rise to tort duties are varied, and so too are the ways in which questions relating to insurance may arise. Our analysis across the course of Chapter 7 and this chapter reflects a spectrum of relationships in which questions of risk allocation arise. Our spectrum is in the reverse order from Stapleton’s analysis,³¹ which begins with stranger cases, and treats this as the paradigm by which other cases are to be judged. At one end of our spectrum are those cases where the two parties themselves are in a contractual relationship,³² or a relationship akin to contract.³³ Close to these are other cases where there is no privity of contract between the parties, but their arrangements are linked through a contractual risk structure.³⁴ In these circumstances, the parties’ risk arrangements have been understood to affect tort duties, or to affect the courts’ reasoning as to the ambit of such duties; but the role of insurance has been underplayed. These cases, beyond concurrency, have been more or less firmly understood in terms of the ‘contractual matrix’ within which the parties’ relationship is to be considered; but the significant link with contextual analysis of contracting, and the need to soften tort’s own two-party frame, has been much less securely acknowledged; and the role of insurance least recognized of all.

Further along the spectrum of tort duties, there are cases where party arrangements are less obviously relevant, or where only one of the litigants in the tort claim is party to a contract purporting to allocate risks.³⁵ In these areas of law, the challenge is to show how courts are still influenced in their approach to tort duties by awareness that they are allocating risks; and to consider what role the pattern of insurance arrangements (rather than the fortuitous presence of insurance) plays in determining such duties. This may mean that structuring of risks by parties remains significant. Or it may mean that society’s preferred response to risks can be discerned through statutory and other arrangements, and that tort duties are thought to destabilize these arrangements, or to be unnecessary. A key point is that even at the ‘traditional’ end of the spectrum there is a concentration of claims in a few categories, including claims against public authorities (though these are varied in their nature); claims against employers (which, of course, is not an instance of claims between ‘strangers’); and claims in relation to road traffic accidents. Cases of the former sort have tested the appellate courts in many jurisdictions, as is well known, and some are explored here. In the latter two instances, as in a number of others, liability insurance is compulsory. The influence of insurance is inclined to be quite different in such cases, but is still not straightforward. We explore the nature of that influence in Chapter 9. But the recognition that there are risk structures in place far beyond contractual matrix cases has also enabled questions of risk allocation to be patently discussed in relation to the duty of care even in some personal injury actions. It is perhaps in relation to personal injury that the reversal of an expected trend to increased welfarism seems most stark. But these cases show that the reasons for the restraint of tort may as readily be products of perceived social preferences about the allocation of risk in pursuit of social goals, as of party arrangements.³⁶ The complex mixing of, and cross-fertilization between, varied social and private goals, legal forms, and associated working concepts (one could also say, ‘ideologies’)³⁷ in relation to supposedly ‘traditional’ tort claims is addressed in detail in Chapter 9, with a particular focus on compulsory liability insurance.

Uniting these examples is the observation that tort duties are disciplined—and increasingly so—by reference to other forms of ordering of risks. This can be seen in different ways across the spectrum of cases, leading from concurrency, through contractual matrix cases, to ‘stranger’ or ‘direct physical injury’ cases of one sort or another, which occur with regularity in respect of particular risks where insurance implications are well recognized. Discounting this can lead to the undermining of settled or preferred risk structures, particularly where the identity of insurers themselves as litigants is thereby overlooked.

During an important, restrictive phase in the evolution of tort duties, John Fleming proposed that the extension of tort into the domain of contract had worked in two directions. As had been predicted, contract had been socialized, absorbing some of tort’s facility with the imposition of duties.³⁸ But at the same time, tort had ‘learned to respect the important message of private ordering’.³⁹ In other words, tort has been part of the same movement noted in Chapter 7, in which the law of contract has developed a contextual approach to assessment of the parties’ allocation of risks in light of their arrangements as a whole.⁴⁰ According to Fleming, this sense of discipline in the law of tort arose in line with the shift from expansionist and generalizing, to restrictive approaches to the ambit of the law of negligence in particular.⁴¹ We agree; and we find the same pattern of restriction, and of concentration on ex ante ordering (private or not), still reflected far beyond the recognized ‘contractual matrix’ cases. Some such cases may reflect ‘private ordering’; others may reflect alternative, publicly mandated, or preferred responses to risk. This process, by which tort duties are restrained in the light of ex ante ordering of risk, and most particularly the role of insurance within that process, are the core themes of this chapter.

8.2 Overview: Tort Reasoning and Risk Allocation

8.2.1 Negligence duties

The very idea that tort duties are to be seen as influenced by questions of risk allocation, even without the additional focus on insurance, will appear incompatible with some influential views. It has been tenaciously argued, for example, that a ‘duty of care’ must be understood as simply a duty to be careful. Courts therefore cannot ‘know’ what the extent of the duty is or what risks fall within it (or ‘which consequences of its breach were those to which the imposition of that duty was directed’), for its meaning is simply to require a certain standard of conduct.⁴² Associated with this is a view that tort duties essentially are owed to the world in general, and simply specify how parties should conduct themselves (‘act with due care’; ‘do not defame’). We disagree with this reading of tort duties. The portrayal of a duty of care as solely an obligation to be careful has been deployed in order to resist analysis of what have been more typically called ‘remoteness’ issues in terms of the ‘scope’ of particular duties.⁴³ Our core concerns are different, but the assertion is nevertheless related to our themes, because the recognition of tort duties, on our account, allocates particular risks, often between particular parties, and does so with due regard to existing risk structures.⁴⁴

The specific and relational nature of tort duties are important elements of our account of their role in allocating risks: if negligence duties did simply give instructions to be careful,⁴⁵ the duty concept could have no risk-allocation function. Stapleton has argued, in support of the idea that negligence duties are merely duties to act carefully, that the denial of a duty to some claimants but not others is ‘discriminatory’ and ‘distasteful’ because it suggests that the defendant is ‘allowed’ to act negligently towards those who are not owed a duty.⁴⁶ We leave aside our doubts about whether this is really so when compared with alternative avenues for defining the limits of negligence liability,⁴⁷ and further doubts as to whether duty-owers are really ‘allowed’ by the law not to act carefully towards certain people simply because they do not owe tort duties in respect of particular risks. More fundamentally, we suggest the courts use the idea of ‘duty of care’ to mean a duty to take care of certain interests. This is captured in the very idea of being careless ‘towards’ someone. The duty is owed to particular people, in respect of particular interests, and is a duty to take care of their interests. Boundaries are therefore inherently important to any statement of a tort duty, and are not merely to be seen as constraints upon those duties.

That duties are not simply defined by their content is a clear implication of the approach in Caparo v Dickman.⁴⁸ That approach is too well established to be dismissed as a consequence of one particular decision, or a handful of decisions. For example, it is clear that the duty of care is owed only to some people, not to others;⁴⁹ only for some purposes;⁵⁰ and only in respect of certain harm.⁵¹ Stapleton acknowledges that the first of these observations is correct.⁵² But she continues to maintain that the ex ante role of negligence duties in particular is simply to offer guidance to act (here, to audit accounts) with due care. How can the duty be limited to particular recipients (as in Caparo) while its ex ante effect (simply to require due care) is said to be so general? Stapleton acknowledges constraining factors, which limit the duties owed by a careless defendant, but prefers to treat them all as ex post ‘control factors’. Is this simply an alternative route to the same end? For the reasons set out in this chapter, we think that it is important to reject the idea that tort duties are merely guidance directives, and to accept the portrayal of tort duties as relational.

Defamation can be offered as an illustration. Stapleton argues that Goldberg and Zipursky’s argument has ‘insuperable problems’ with torts other than negligence, and illustrates this with reference to defamation. We are not convinced that the ‘relational’ approach faces insuperable problems accounting for defamation. On the contrary, defamation torts depend on statements made of an individual or a limited class in which the individual is identifiable, and are not best understood as duties owed to the world. Defamation duties are at least as ‘relational’ as negligence duties: they are duties not to publish statements defamatory of X or Y. They are not easily understood as duties ‘not to make defamatory statements’ in the abstract, as the idea of a defamatory meaning or statement cannot exist without reference to an affected party. Defamation underlines the relational nature of tort duties. Similar points may be made about other torts, such as deceit: there must be an intention to mislead the claimant specifically, in such a way that they act in reliance. The relational approach does not struggle to account for torts other than negligence.

Equally, we do not accept that strict duties in general demand any different analysis. It may certainly be doubted whether it is true that there are no strict tort duties to succeed,⁵³ even if statutory provisions are for some reason excluded from the ambit of ‘tort’. Apart from the possibility that negligence imposes such duties,⁵⁴ it is arguable that there is a strict duty to succeed in not falsely imprisoning someone, for example, in that an imprisonment which is false remains false even if the defendant was reasonable in his or her belief that it was lawful: R v Governor of Brockhill Prison, ex p. Evans.⁵⁵ That trespass torts (such as false imprisonment) are viewed as requiring a breach of duty is now settled in English law, at least for the purposes of interpreting provisions of the Limitation Act 1980, and it may therefore be concluded that here is not only a strict liability, but a clear instance of a strict duty to succeed.⁵⁶

It is consistent with a relational view of tort duties to propose that to be owed a duty of care is for the risk of negligence within the ambit of the duty to be placed with the person who owes the duty, so far as a pattern of legal rights and duties can do it.⁵⁷ The ‘risk allocation’ view of tort law does not collapse into a denial of duty or a reduction of duty to mean only the risk of liability.⁵⁸ Put simply, courts and legislatures take seriously both the nature of obligation (tort duties must be consistent with the sense of being under a duty, even if it is a strict duty to avoid a particular result); and the impact of liability which is a necessary corollary of recognizing or imposing a tort duty. Across the spectrum of tort, duties are determined on the basis of a wide variety of reasons (or criteria). At the more restrictive end of the negligence spectrum, it can be argued that duties are imposed where it is concluded that the risk was one that the claimant could treat as taken care of by the defendant.⁵⁹ Across the spectrum, not merely in exceptional cases, the impact of liability is reflected in the definition of tort duties. To think in terms of risk allocation is not to negate duties, nor even to subject them to policy restraint, but is an important element in the reasoning process behind the creation of those duties as a matter of law.

8.2.2 Evolving styles of reasoning

Decisions of the courts will express the issues just debated to different degrees, but risk allocation questions are no longer hidden beneath the surface of judgments. Equally, it has become more difficult to deny that insurance is relevant to judicial decision-making as it is discussed as a factor in an increasing number of instances and, as we show throughout, at pivotal moments in the law’s development.⁶⁰ This partly reflects the more general way in which judicial reasoning styles have gradually evolved where tort duties are concerned. A formal style of reasoning can be recognized in which relatively abstract categories such as ‘foreseeability’, ‘proximity’, or more detailed ideas such as ‘control’, are applied to the case in hand. English courts have in the past been criticized for adopting a formalistic approach along these lines, and for not grappling with policy concerns, or failing to make plain what the motivating reasons were.⁶¹ It is our suggestion that English courts now adopt a more pragmatic style of reasoning around tort duties, despite the existence of umbrella terms such as ‘proximity’. In particular, courts have drawn attention more confidently to the impact of allocating risks to defendants through imposing a duty.

We should clarify that our point about pragmatism does not amount to a general argument about the incorporation of ‘policy considerations’ in all their multiple guises. Indeed, if ‘policy’ is chiefly concerned with considerations beyond the parties, some of the pragmatism to which we refer is not properly described in terms of policy at all (unless a particularly hard-line bipartite approach is adopted to the question of what ‘relates to the parties’ and their relationship). This reflects the fact that in relation to insurance, we do not think that courts are primarily reaching for ‘community welfare’ in a direct sense,⁶² although they may of course be influenced by the existence of legislative structures which reflect judgments about community welfare.

The issues here go back once again to changing understandings of the context in which the law of obligations operates, and of the social and economic role of this aspect of private law. We have already touched here on a shift from welfarism (in which tort trumps contract) to market contextualism, in which the significance of party intentions is reinstated and the law’s role becomes more reflexive, encouraging and supporting party self-regulation. In terms of the law of tort, this shift was adeptly captured by Sir Anthony Mason, writing the Preface to the 9th edition of Fleming’s Law of Torts.⁶³ He fittingly observed a reorientation from ‘economic “welfarism”’ to ‘economic “rationalism”’, which he perceived as a vindication of John Fleming’s scholarship. He described this reorientation as strikingly illustrated by Stovin v Wise,⁶⁴ a decision of the House of Lords in which attempted risk-shifting by insurers was, as we explain below, centrally implicated.⁶⁵ We explain why we consider this an astute depiction, in 8.4. This trend has continued. The most recent edition of Fleming’s The Law of Torts continues the theme, containing a Foreword⁶⁶ which makes express the transition made by the English courts from formalism to overt recognition of the implications of liability. Lord Hoffmann describes a ‘different style of judgment-writing, in which judges are more open about the real reasons for their decisions and less inclined to think that the authority of their priesthood requires concealment of the inner mysteries of the craft’. This further encourages our view that risk-allocation questions are among the ‘real reasons’ for decisions, and are not to be dismissed as mere ‘makeweight’ arguments.⁶⁷

Quite possibly, the road map to duty of care questions offered by Caparo v Dickman has turned out to be better adapted to certain forms of pragmatism than the rival approaches of Australian, Canadian, and New Zealand courts, as they have diverged from the English approach in relation to duties of care, despite (or because of) their more open embrace of wide-ranging policy concerns. Tort rarely, if ever, represents the only response to risk in operation,⁶⁸ and this is increasingly recognized by the pragmatic style articulated by the House of Lords in Caparo. As concern with the impact of expansive tort liability in all contexts has mounted, English courts have offered more express justifications for the restrained shape of liability. These reasons are, quite rightly, not necessarily labelled ‘policy’. In this process, insurance concerns, among others, have been at work in restricting, rather than only expanding,⁶⁹ liabilities, because insurance arrangements often reflect existing forms of ordering, whether public or private.

In the remainder of the chapter, we begin our analysis with ‘contractual matrix’ cases, and explore the role of insurance in relation to party risk structures (8.3). We then extend our examination into the area of advice (8.4), and beyond that to the varied category of local authority liability (8.5). These are among the recognized pressure points for delineation of tort duties. As we move along the spectrum, risk allocation increasingly becomes one of a variety of considerations; and yet we suggest that it does not become irrelevant. Deferring our analysis of the non-paradigmatic paradigm—running-down cases—for the next chapter, we nevertheless consider one area where duty relationships are both well recognized and backed by compulsory insurance, and where the structure of duties is complex. This is the area of employers’ duties to employees (8.6), where we briefly examine the structure of duty and the roles of insurance in order to give a sense of the variety and depth of influence of insurance in relation to tort duties.

8.3 ‘Contractual Matrix’ Cases: Tort and Party Risk Structure

It has been widely recognized that tort duties may be embedded within market arrangements, raising pertinent questions about the nature of those duties and their relationship with the law of contract. On the tort side, such issues have been approached in terms of a category of ‘contractual matrix’ cases, where parties ‘come together against a background of contract’, or where the risks involved are generally planned, even where the parties to the tort action do not themselves contract directly with one another.⁷⁰ Initially, the issues were thought relevant only to a special category of ‘pure economic losses’. The key point however is that the parties make provision for the economic cost of various harms, so that it is really the pattern of relationships and of risk allocations which offers these cases their complexity, not the category of loss involved. Referring in particular to recovery in tort for purely ‘economic’ losses, such cases have been described as raising particular issues precisely because they are more likely than other cases to be associated with ‘planned transactions’, and ‘contractual expectations’.⁷¹ More recently, it has become much clearer that these factors may also be present in cases of physical damage.⁷²

Use of the term ‘contractual matrix’ in this context is most associated with John Fleming,⁷³ who drew it from the judgment of La Forest J in the Supreme Court of Canada in London Drugs v Kuehne & Nagel.⁷⁴ Here, the Supreme Court held that a contracting party who had agreed to a limitation in liability in respect of damage to its goods could not evade this contractual limitation by proceeding in tort against an employee of the company with whom it had contracted. The limitation of liability operated also for the benefit of the employee. This has been treated as a significant relaxation in the privity requirement in contract, and has subsequently been deployed in different circumstances by the same court.⁷⁵ La Forest J, agreeing in the result, dissented on the separate question of whether the employee owed any duty of care at all: in his view, no duty of care arose in the context of the particular allocation of risk observable in the contractual matrix. It appears that the claimant had taken the obvious course when faced with a limitation of liability—indeed, the expected course—and insured against loss above the contractual limit. The action was brought by the claimant’s insurer, and the approach of La Forest J is therefore in accordance with our analysis in Chapter 5, namely, his approach to relevant duties identified the insurer as the primary indemnifier against the background of the contractual structure or matrix. La Forest in turn referred to the reasoning of the English Court of Appeal in Pacific Associates v Baxter as exemplifying the influence of the ‘contractual matrix’ in tort reasoning.⁷⁶

It is not an accident that the expression ‘contractual matrix’ recalls the idea of the ‘factual matrix’ which recurs in contextual analysis of contract terms and their meaning.⁷⁷ Like the ‘new contextualism’ which has become orthodox in relation to the English law of contract, the logic of analysis of tort duties against the background of a ‘contractual matrix’ seeks a purposive interpretation of the risk structure adopted by the parties. That being the case, it is not really to the point that in some (though by no means all) instances, literal agreement to waive or exclude rights in tort cannot be found: intentions may legitimately be inferred from existing patterns of risk allocation, even if this is avowedly an interpretive exercise. This is inherent to the difference between literalism and contextualism.⁷⁸ In this light, Stapleton’s doubts about the expression ‘contractual matrix’, and preference for the terminology of unbroken contractual ‘chains’, is itself significant.⁷⁹ We suggest that this terminological preference reflects unease with the idea of multi-party structuring of risks which is also reflected in reluctance to allow risk structures falling short of contractual terms to override the perceived functions of tort duties (in Stapleton’s account, primarily deterrence).

Even among those existing analyses of ‘contractual matrix’ cases which share some ground with our position, looking at liability and associated risks in ex ante terms, insurance has not been fully absorbed within the process, despite the significant role demonstrated for insurance in Chapter 7. Some have particularly resisted the softening of tort’s focus on two parties, and the implication that responsibility and deterrence in relation to careless behaviour do not necessarily trump the allocation of risks associated with that behaviour, simply by turning from contract to tort.

In particular, while proposing that the ability of claimants to achieve alternative protection than from the defendant is a key factor in determining duty questions, Jane Stapleton has forcefully limited the range of acceptable opportunities, to encompass only those which are consistent with enhanced deterrence. It is on this basis that she claims that insurance in particular cannot be a consideration.⁸⁰ This rests on a strict, verging on absolute, interpretation of the nature of the law of tort as aimed at achieving deterrence, which is deployed particularly forcefully at the exclusion of insurance. Stapleton argued that the idea of alternative opportunities of protection open to the claimant is too crude a concept to be relied upon in the limitation of tort duties without refinement. For example, ‘it could not refer to any means of self-protection for it could not mean self-protection by the purchase of first-party insurance’. The footnote to this statement continues, ‘In deciding whether to protect a plaintiff the law of tort does not take his or her insurability into account, nor could it do so and also seek coherently to promote deterrence of defendant carelessness’.⁸¹