-- To obtain permission to use this article beyond the scope of your HeinOnline license, please use:

Transcription

1 Citation: 55 Cambridge L.J Content downloaded/printed from HeinOnline Fri Apr 21 04:25: Your use of this HeinOnline PDF indicates your acceptance of HeinOnline's Terms and Conditions of the license agreement available at -- The search text of this PDF is generated from uncorrected OCR text. -- To obtain permission to use this article beyond the scope of your HeinOnline license, please use: Copyright Information

2 Cambridge Law Journal, 55(3), November 1996, pp Printed in Great Britain REMOTENESS OF DAMAGE IN CONTRACT AND TORT: A RECONSIDERATION JOHN CARTWRIGHT* THE starting point for any rule of remoteness of damage is the familiar notion that a line must be drawn somewhere: it would be unacceptably harsh for every tortfeasor or contract breaker to be responsible for all the consequences which he has caused. We could, of course, just abandon any attempt to formulate rules for this area, and look to the judge's pragmatic discretion as to what is "fair" in the circumstances of the particular case: let it be simply a question of fact. Certainly, the question of where to draw the line on recoverability of consequential losses cannot be answered by a mathematically precise formula, and occasionally the judges are quite explicit that they are drawing the line on the facts of a case because they regard that as a "fair" solution.' Yet considerations taken into account by the judges in deciding particular cases can and should be identified: even the notion of "fairness" has to be judged from some particular perspective. The possible policies which can be reflected in the choice of a remoteness rule will therefore include those which view the problems from the perspective of either the plaintiff or the defendant, as well as those which are in a sense neutral to the two parties. Various possible policies can be identified. One approach would take account of the nature of the harm suffered by the plaintiffand so we might distinguish between injuries to the person, to property and to economic interests. If, say, the plaintiffs interest in his physical integrity is classified as being of a higher nature than his purely economic interests, the threshold for recovery of damages for physical injuries would be lower. We shall see that this distinction has occasionally been used in the context of remoteness of damage, 2 although it is has not gained acceptance as a test in its own right. 3 Tutor in Law, Christ Church, Oxford. Mr. J.W. Davies made many helpful and perceptive comments on an earlier draft of this article, but (of course) bears no 'responsibility for the final text. E.g. Watkins L.J. in Lamb v. Camden LBC [1981] Q.B. 625, Parsons v. Uttley Ingham [1978] 1 Q.B. 791, ; Hart and Honor6, Causation in the Law (2nd ed., Oxford 1985), pp ; Cooke [1978] C.L.J. 288, 298. In Cambridge Water Co. v. Eastern Counties Leather plc [1994] 2 A.C. 264, 300 Lord Goff argued that a plaintiff should not be able to recover for damage to property more easily than personal injury. Ibid., at pp. 804, 806.

3 C.L.J. Remoteness of Damage in Contract and Tort 489 A different principle might link the recoverable losses to the defendant's culpability. Instead of asking, "why should the plaintiff be able to recover?", this would ask, "how far is it fair to make the defendant liable?" It would be rational to have a general rule (in both tort and contract) that a defendant who is a deliberate wrongdoer should be treated differently from one who is only negligent or inadvertent; and this is in fact adopted in some torts, but not in the contract remoteness rule. Economic arguments are, in a sense, more neutral: these might ask which party is better placed to bear the losses which flow from the wrongful act-and, for example, which party can insure against such losses (either the more efficiently or at all). 4 This approach would also consider whether the choice of a particular remoteness rule would promote more desirable policy objectives. So, it is sometimes argued that a remoteness rule in contract should be designed to promote the efficient disclosure of information about possible losses by parties during the negotiations leading up to the contract. 5 A broader and more pervasive approach, however, is to look to the interest protected by the duty of which breach has given rise to the claim. The damage recoverable is inextricably linked to the cause of action in respect of which it arises. This is a rational approach which expects consistency and integrity within the body of rules governing a claim for breach of contract, or a particular tort. The various approaches and policies can, of course, be combined; but the judges have to start somewhere: it is not surprising that a "default" rule has developed-a rule which will be applied unless there are identifiable characteristics of the wrong concerned, or the particular facts (such as the way in which the wrong was committed, or wider (but definable) issues of policy) which justify departing from it. It would be possible to use as the default rule that all losses suffered by the plaintiff in consequence of the breach of duty are recoverable. 6 But the English courts have adopted a different position. In The Wagon Mound 7 Viscount Simonds took the view that [i]t is a principle of civil liability, subject only to qualifications which have no present relevance, that a man must be considered to be responsible for the probable consequences of his act. To demand more of him is too harsh a rule, to demand less is to ignore that civilised order requires the observance of a minimum standard of behaviour. These words were spoken in a case discussing remoteness of damage Lamb v. Camden LBC, above, n. 1, at pp Posner, Economic Analysis of Law (4th ed., Boston 1992), p. 127; Bishop (1983) 12 J.L.S. 241, ; Cane, Tort Law and Economic Interests (2nd ed., Oxford 1996), p Cooke [1978] C.L.J [1961] A.C. 388,

4 The Cambridge Law Journal [1996] in negligence, but they are general and appear to reflect the idea that the "default" rule is that remoteness is limited by reference to the foreseeable consequences of one's acts. This is a position based explicitly on a moral view of the law of obligations, and takes the starting-point as being from the defendant's position: we should normally limit the defendant's liability to that which he could have foreseen. There has been a tendency in recent years to move firmly in the direction of this as the default position. It has been adopted in relation to nuisance (both public and private) and the rule in Rylands v. Fletcher;' and defamation. 9 In order to examine the whole issue, it will be convenient to begin by concentrating on a specific area, the rules for contract, before comparing the tort of negligence, as well as (briefly) other torts. THE CONTRACT RULES FOR REMOTENESS OF DAMAGE The starting point here must be The Heron II, since it contains an explanation of the content and rationale of the rule governing remoteness of damage in contract. The House of Lords made clear, first, that there is a general remoteness rule in contract based on what the defendant could have foreseen at the time of the contract; and second, that this rule requires a relatively high level of foreseeability or "contemplation", which is (in particular) higher than the equivalent test in tort. "Contemplation" at the Time of the Contract It has been clear at least since Hadley v. Baxendale that the contract rule is based on a degree of foreseeability at the moment of entering into the contract. The rule laid down by Alderson B." was that damages are such as may fairly and reasonably be considered either arising naturally, i.e., according to the usual course of things, from [the] breach of contract itself, or such as may reasonably be supposed to have been in the contemplation of both parties, at the time they made the contract, as the probable result of it. Alderson B. explained the reason for such a rule: if the plaintiff could recover losses arising out of circumstances which the defendant could not have had in contemplation at the time of the contract, this would impose on the defendant too harsh a liability. If there are such "special The Wagon Mound (No. 2) [1967] A.C. 617; Cambridge Water Co. v. Eastern Counties Leather pic, above, n. 2. Slipper v. BBC [1991] 1 Q.B '0 [1969] 1 A.C " (1854) 9 Ex. 341, 354.

5 C.L.J. Remoteness of Damage in Contract and Tort circumstances", the plaintiff will not be able to recover losses flowing from them unless he makes the knowledge common by telling the defendant:12 Now, if the special circumstances under which the contract was actually made were communicated by the plaintiffs to the defendants, and thus known to both parties, the damages resulting from the breach of such a contract, which they would reasonably contemplate, would be the amount of injury which would ordinarily follow from a breach of contract under these special circumstances so known and communicated. But, on the other hand, if these special circumstances were wholly unknown to the party breaking the contract, he, at the most, could only be supposed to have had in his contemplation the amount of injury which would arise generally, and in the great multitude of cases not affected by any special circumstances, from such a breach of contract. For, had the special circumstances been known, the parties might have specially provided for the breach of contract by special terms as to the damages in that case; and of this advantage it would be very unjust to deprive them. Alderson B. referred to "the parties" being deprived of the chance to bargain in the light of the special circumstances, but the rule is designed primarily with the defendant's position in mind: the underlying idea is that the defendant should not be responsible beyond that which he could have realised; and this realisation is tested at the time of the contract (rather than, for example, breach) because once he is bound to the contract it is too late for him to negotiate any other terms into the contract which might balance the "special" risk which he is undertaking. 3 The rule is not devised exclusively from the defendant's perspective: the plaintiff is entitled to assume that the defendant has knowledge of "ordinary" risks, and so need not draw them to his attention. In a sense, it balances the interests of the parties, although the main thrust is to view the problem from the defendant's position. The rule can be argued to be efficient in that it encourages disclosure by the plaintiff of information which is not otherwise available to the defendant, but which is relevant to assessing the risks of the plaintiff's loss in the event of breach. 4 That is not to say, though, that the judges who have developed this rule have done so explicitly for reasons of economic efficiency. The underlying logic of Hadley v. Baxendale and the cases which follow it is that, to be fair to the defendant, he should be responsible only to the extent of the risks which he could have realised 2 Ibid., at p ' See also Victoria Laundry (Windsor) Ltd v. Newman Industries Ltd. [1949] 2 K.B. 528, 539; The Heron II, above, n. 10, at pp. 386, 397, Bishop (1983) 12 J.L.S. 241,

6 The Cambridge Law Journal [1996] he was undertaking when he agreed to the contract, and so the risk of loss is one which he can be held to have undertaken. From time to time, it has been said 5 that the remoteness test in contract is based on an acceptance by the defendant of the risks of the plaintiff's losses. It is clear, though, that the test does not require a contractual promise (express or implied) by the defendant about the extent of his liability; nor is any express acceptance of risk required. The best explanation was given by Bridge L.J. in GKN Centrax Gears Ltd v. Matbro Ltd :16 If a party entering into a contract would as a reasonable man foresee that his breach would be likely to cause loss of a certain kind to the other contracting party, then it must be from that and nothing else that the implication would arise of his contractual willingness to accept liability for that loss. The question is therefore what risks the defendant can be taken to have accepted within the contract; and the answer is found by considering what a reasonable man in the circumstances would have realised as the likely consequences of his breach. It is the defendant's entering into the contract in circumstances where he knows or ought to know of these risks, that shows his willingness to bear them. 7 What Degree of "Foreseeability" or "Contemplation"? Comparison with Tort The test for remoteness of damage in contract in The Heron 1I therefore sets the limits of the defendant's responsibility by what he had, or could have had, in contemplation when he committed himself to the contract. The next-and crucial-question is what degree of "foreseeability" or "contemplation" is required.' 8 A discussion of this issue can quickly degenerate into a semantic exercise 9 which purports to embody an analysis of differing degrees of probability of result. If we say that a defendant "reasonably could have foreseen" a particular consequence, how different is that from saying that the result was '5 British Columbia and Vancouver's Island Spar, Lumber and Saw Mill Co. Ltd v. Nettleship (1868) L.R. 3 C.P. 499, 506; GKN Centrax Gears Ltd. v. Matbro Ltd. [1976] 2 Lloyd's Rep. 555, 580; The Pegase [1981] 1 Lloyd's Rep. 175, 182; Seven Seas Properties Ltd. v. Al-Essa (No. 2) [1993] 1 W.L.R. 1083, Above, n See also The Pegase, above, n. 15, at pp IS There is also a question of what must be foreseeable. It appears that the defendant need not foresee the actual losses, only the kind of losses which would be suffered by the plaintiff: Christopher Hill Ltd. v. Ashington Piggeries Ltd [1969] 3 All E.R. 1496; GKN Centrax Gears Ltd v. Matbro Ltd, above, n. 15, at p. 580; Parsons v. Uttley Ingham [1978) 1 Q.B. 791, 813; The Rio Clara [ Lloyd's Rep. 173, 175; or the circumstances in which the losses would be suffered: Great Lakes Steamship Co. v. Maple Leaf Millin Co. Ltd (1924) 20 L. L. Rep. 2, 5, 7; Wroth v. Tyler [1974] Ch. 30, 61. The application of this principle may be more difficult in cases of economic harm than physical damage: see n. 65, below. '9 Parsons v. Uttley Ingham, above, n. 18, at pp. 802, 807.

7 C.L.J. Remoteness of Damage in Contract and Tort "probable", "possible", "a real possibility", "liable to result", "in the defendant's contemplation"? This issue exercised the House of Lords in The Heron II, and it is clear that the House wished to emphasise that, whatever language is used to describe the test in contract, it is not the same as the test in tort: the contract test is narrower, in the sense that a higher degree of probability is necessary in contract than in tort. The question, "how foreseeable?" cannot be separated from the question of why we use a test based on foreseeability: the content of a rule must fulfil the policy of the rule. Given that the purpose of the remoteness rule in contract is to make the defendant liable for only the risks which he had or could have had in contemplation when he was settling the balance of the bargain with the plaintiff, it would be natural for the courts to look to the risks which the parties (but particularly the defendant) could realistically have had in mind as the consequences of his failure to fulfil his obligations. Indeed, in spite of the variation of language used in The Heron II, the House of Lords did require a realistic level of foreseeability by the defendant. Lord Upjohn, for example, said: 2 in contract the parties have only to consider the consequences of a breach to the other; it is fair that the assessment of damages should depend on their assumed common knowledge and contemplation and not on a foreseeable but most unlikely consequence. The parties may moreover agree to limit or exclude liability for damage, or agree on a liquidated sum, or one party can disclose to the other special circumstances which will render a breach especially serious to him. This was contrasted with the position in tort, where the test is wider: the tortfeasor is liable for any damage which he can reasonably foresee may happen as a result of the breach however unlikely it may be, unless it can be brushed aside as far fetched. See the Wagon Mound 2 cases. This point was made by other members of the House, and in most detail by Lord Reid. 22 There was much discussion about the language appropriate to describe the degree of foreseeability in contract and in tort. In particular, Lord Reid emphasised that the phrase "reasonably foreseeable" is too low a threshold for the remoteness test in contract. In saying this, he was criticising a section of the judgment of Asquith L.J. in the Victoria Laundry case which had referred to the test in contract as allowing recovery of losses which were "reasonably 20 [1969] 1 A.C. 350, [1961] A.C. 388; [1967] A.C [1969] 1 A.C , 389; see also Lord Hodson at p. 411 and Lord Pearce at pp

8 The Cambridge Law Journal [19961 foreseeable as liable to result from the breach". This may be a harsh criticism of Asquith L.J., since it can be argued that, read in context, his language did express a realistic level of foreseeability; and there is nothing in what he said to indicate whether he regarded the contract and tort tests as identical. He said that reasonable foreseeability depends on the parties' (or at least the defendant's) knowledge at the time of the contract; and that this means on-hadley v. Baxendale principles-that he could, as a reasonable man, foresee it was likely to result: a serious possibility or a real danger. 23 On this view, the House of Lords in The Heron II were unjustly criticising AsquithL.J.'s use of the words "reasonably foreseeable" as being (1) too low a threshold of probability and, in consequence, (2) a confusion with the tort test: these two matters had been established not by 1949 (when Asquith L.J. gave his judgment in Victoria Laundry-and when, strictly speaking, the tort test was quite different 24 ) but since 1961 (in The Wagon Mound and the cases 5 which followed it in establishing that a loose form of "reasonable foreseeability" is the test in tort). On the other hand, the Privy Council in The Wagon Mound had sown the seeds of the problem which the House of Lords felt it necessary to address in The Heron II, since Viscount Simonds, giving the judgment of the Board, explicitly took the view that contract and tort remoteness principles are the same; and that there was nothing to suggest that Asquith L.J. in Victoria Laundry had not himself thought this too. 2 6 It was therefore natural that, in the first contract remoteness case to come before the House of Lords after the developments in tort in The Wagon Mound cases, the argument should be put that the tort and contract tests were indistinguishable. The difficulty arose from the development of the tort cases which emphasised that relatively low risks of loss are within the tort test. Lord Reid in The Heron II was perhaps particularly sensitive to this issue, since it was he who had given the judgment of the Privy Council in Wagon Mound (No. 2), in which he had said 27 that a question in tort is whether the risk is-one which would occur to 23 [ K.B. 528, (propositions (3.) to (6.)). The only real problem with Asquith L.J.'s formulation may be his final suggestion of the expression "on the cards". This expression is colloquial and unclear: whether it means a low or high probability may depend on the "cards" to which it refers: playing cards (low probability, perhaps even a one in 52 chance: see Lord Reid [1969] 1 A.C. 350, 390) or tarot cards (high probability, presumably, in the eyes of one who believes in the accuracy of fortune-telling cards). It is not clear which cards were originally referred to in this expression: Oxford English Dictionary (2nd ed.), Card, sb 1.2.e. 24 Re Polemis [1921] 3 K.B In particular, Hughes v. Lord Advocate [1963] A.C. 837 and The Wagon Mound (No. 2) [1967] A.C [1961] A.C. 388, 420. Nor did the House in East Ham Corporation v. Bernard Sunley & Sons Ltd. [19661 A.C. 406 react to counsel's suggestion (at p. 418) that, after Wagon Mound, "the test of foreseeability is applicable in both tort and contract"; and Lord Upjohn (at p. 445) and Lord Pearson (at p. 451) even used without criticism Asquith L.J.'s test of"on the cards" from Victoria Laundry. 27 Above, n. 25, at p. 643.

9 C.L.J. Remoteness of Damage in Contract and Tort 495 the mind of a reasonable man and which he would not brush aside as far-fetched. Rather than dwell on whether the judges in these earlier cases thought or should have thought that there were differences of substance between the contract and tort remoteness tests, it is more helpful to look to some underlying principles which can be deduced from the cases and which can be used to point to how the courts ought today to deal with the problem. Clearly the Privy Council and the House of Lords had in mind a general principle that a defendant should normally be responsible only for the foreseeable consequences of his act. This is the "default" rule which has already been discussed above, and is most clearly shown by the judgment of Viscount Simonds in The Wagon Mound 8 where he mentioned Hadley v. Baxendale in his argument that there is a common principle in tort and contract, that a person is responsible "for the probable consequences of his act". However, the thinking of the House of Lords in The Heron II is particularly significant. For they adopted an approach based on the underlying purposes of the particular area of law in question. The decision there was, in essence, that the contract remoteness test is set quite differently from the tort test, and for different reasons of principle which are derived from the underlying purposes of contract law: it is because the underlying reasons for the remoteness test in tort are different, that the test itself is different. And the key difference which the House identified was that the relationship between the parties to a tort is different from that between contracting parties. Lord Pearce put it clearly: it was suggested in argument that there was or should be one principle of damages for both contract and tort and that guidance for one could be obtained from the other. I do not find such a comparison helpful. In the case of contract two parties, usually with some knowledge of one another, deliberately undertake mutual duties. They have the opportunity to define clearly in respect of what they shall and shall not be liable. The law has to say what shall be the boundaries of their liability where this is not expressed, defining that boundary in relation to what has been expressed and implied. In tort, two persons, usually unknown to one another, find that the acts or utterances of one have collided with the rights of the other, and the court has to define what is the liability for the ensuing damage, whether it shall be shared, and how far it extends. If one tries to find a concept of damages which will fit both these different problems there is a danger of distorting the rules to accommodate one or the other and of producing a rule that is satisfactory for neither. 28 Above, n. 26, at pp. 419, Ibid., at p See also Lord Reid at p. 386, Lord Hodson at p. 411 and Lord Upjohn at p. 422.

10 The Cambridge Law Journal [1996] But is this rationale for the distinction between the tort and contract remoteness tests good for all cases? The logic is that, where the basis of liability is what the defendant has agreed, then the extent of his liability ought also to be linked to what he agreed, or (given that such matters are rarely expressly stated by the parties) can be taken to have agreed. This leads to a rule of remoteness of damage in contract based on the defendant's "contemplation"-a realistic level of foreseeability. By contrast, the House of Lords assumed that the circumstances in which torts are committed are fundamentally different: parties who meet for the first time on the occasion when one collides with the other. Although this may be the paradigm case of tort, not all cases of tort-even of the tort of negligence-fit this rather sweeping generalisation. THE TORT OF NEGLIGENCE Traditionally, we say that the elements of the tort of negligence are a duty of care; breach of the duty; and resulting damage. Judges have on numerous occasions commented that the issues of duty and remoteness are just aspects of a single question, and that the division into "duty-breach---damage" is simply an analytical tool to determine whether, on the facts of a particular case, a defendant ought to be made to pay for the plaintiff's loss. 3 " However, even if this rather pragmatic line can legitimately be taken, one would expect some link between the tests of "duty" and "remoteness". Indeed, the leading modern case on the test for remoteness of damage in the tort of negligence, The Wagon Mound, drew the line for recovery at damage which "is of such a kind as the reasonable man should have foreseen" for the very reason that the limit of liability in negligence, as set out by Lord Atkin in Donoghue v. Stevenson, 3 is fixed by reference to the reasonably foreseeable consequences of the defendant's 'acts. As Viscount Simonds said, 32 If, as admittedly it is, B's liability (culpability) depends on the reasonable foreseeability of the consequent damage, how is that to be determined except by the foreseeability of the damage which in fact happened? The Privy Council were clearly unhappy with the notion that a defendant whose duty of care rested on what he could reasonably have 3 Lord Pearson in Home Office v. Dorset Yacht Co. Ltd. [1970] A.C. 1004, 1052; Denning L.J. in Roe v. Minister of Health [1954] 2 Q.B. 66, 85 and (as Lord Denning M.R.) Spartan Steel v. Martin [1973] Q.B. 27, 37; and (significantly, given his later statements in Caparo v. Dickman) Oliver L.J. in P. Perl (Exporters) Ltd. v. Camden LBC [1984] Q.B. 342, [19321 A.C. 562, Above, n. 26, at p. 425.

11 C.L.J. Remoteness of Damage in Contract and Tort foreseen should then be responsible for unforeseeable consequences of his breach of that duty. And, if the existence of a duty of care did follow from the defendant's reasonable foreseeability of damage, it would of course make little difference whether one approached a damages claim as an issue of "duty" or "remoteness": in either case, the question would be identical: "could the defendant reasonably have foreseen the plaintiff's damage?" 33 Matters are not so simple. In particular, since 1961 there has been very significant development of the law governing the imposition of a duty of care. It is therefore to be expected that the remoteness test should also itself be re-appraised. "Reasonable Foreseeability" in the Current Approach to the Existence of a Duty of Care There has been much litigation in recent years about the existence and scope of the duty of care. Although the House of Lords have further developed particular areas, the current approach is embodied in the decision in Caparo Industries plc v. Dickman, 34 where the House adopted the approach advocated in the High Court of Australia by Brennan J. 35 : novel categories of negligence should be developed incrementally and by analogy with established categories. Whether we have really dispensed with a generalised test may be questioned: it is becoming clear that, in answering the question, "is this (novel) case analogous to any existing category of duty?" judges are still using what appears to be a three stage test of (1) foreseeability; (2) proximity; (3) justice and reasonableness in imposing a duty. Indeed, members of the House of Lords have done this expressly. 36 But it is certainly clear that the judges are now thinking in terms of duties being defined by reference to different categories of situation. Although it is clear from Caparo v. Dickman that "reasonable foreseeability" is not a sufficient criterion for the existence of a duty of care in all situations, it is none the less clearly a necessary criterion. Some guidance is given in Caparo-and in particular in Lord Oliver's speech 37 -on how to identify and distinguish different categories for which duties of care exist. In particular, criteria relevant to identify a category will include the activity concerned (act? words? omission to act? omission to speak?) the type of loss suffered by the plaintiff 33 Even in 1987 the Court of Appeal were able to take this view in the context of liability for nervous shock: Attia v. British Gas plc [1988] Q.B. 304, 315, [ A.C. 605, , 635, 643, rejecting the idea of a generalised test of Anns v. Merton LBC [1978] A.C. 728, Sutherland Shire Council v. Heyman (1985) 60 A.L.R. 1, Spring v. Guardian Assurance plc [1995] 2 A.C. 296, 334, 342, 345; White v. Jones [1995] 2 A.C. 207, ; Marc Rich & Co. AG v. Bishop Rock Marine Co. Ltd. [1996] 1 A.C. 211, 235, Above, n. 34, at pp See also Lord Bridge at p. 618 and Lord Roskill at p. 629.

12 The Cambridge Law Journal [1996] (bodily injury? psychological injury? property damage? economic loss?) and the link between these two (direct infliction? indirect infliction either with, or without, the intervention of third paries?). Caparo v. Dickman involved the infliction of economic loss on the plaintiff directly by his reliance upon a statement issued by the defendant, and the case stands as authority that a duty of care is owed in such situations where particular requirements are satisfied: in essence, that the defendant should reasonably have foreseen that the plaintiff would be a recipient of the statement and would probably rely upon it for a particular purpose. 8 The detail of the rule for this category is not however our main concern. What is important is to notice that foreseeability of the particular kind of damage suffered by the plaintiff is an indispensable element in the search for the existence of the duty of care. Lord Bridge 39 was explicit about this: It is never sufficient to ask simply whether A owes B a duty of care. It is always necessary to determine the scope of the duty by reference to the kind of damage from which A must take care to save B harmless. This idea draws on the judgment of Brennan J. in Sutherland Shire Council v. Heyman, who said: 4 " Liability in tort is for damage done, not for damage merely foreseeable or threatened or imminent. The principle was stated by Viscount Simonds in delivering the judgment of the Judicial Committee in The Wagon Mound: 'It is, no doubt, proper when considering tortious liability for negligence to analyse its elements and to say that the plaintiff must prove a duty owed to him by the defendant, a breach of that duty by the defendant, and consequent damage. But there can be no liability until the damage has been done. It is not the act but the consequences on which tortious liability is founded... It is vain to isolate the liability from its context and to say that B is or is not liable, and then to ask for what damage he is liable. For his liability is in respect of that damage and no other. If, as admittedly it is, B's liability (culpability) depends on the reasonable foreseeability of the consequent damage, how is that to be determined except by the foreseeability of the damage which in fact happened-the damage in suit?' The corollary is that a postulated duty of care must be stated in reference to the kind of damage that a plaintiff has suffered and in reference to the plaintiff or a class of which the plaintiff is a member... It is impermissible to postulate a duty of care to avoid one kind of damage-say, personal injury-and, finding 3 Ibid., at pp. 621, 638, Ibid., at p See also Lord Oliver at p (1985) 60 A.L.R. 1, 48, quoted by Lord Oliver at [ A.C [1961] A.C. 388, 425.

13 C.L.J. Remoteness of Damage in Contract and Tort the defendant guilty of failing to discharge that.duty, to hold him liable for the damage actually suffered that is of another independent kind-say, economic loss. The issue is here illustrated very neatly: by reference to the case which is generally seen as the leading case on remoteness of damage in negligence, Brennan J. emphasised that the tests for both the duty of care and remoteness of damage hinge on the same idea: that a defendant's liability (duty) is in respect of a particular kind of damage suffered by the plaintiff. Duties are owed in respect of particular kinds of damage, and the court cannot decide whether a duty was owed without asking whether the defendant could reasonably have foreseen that carelessness on his part might result in damage of the kind which the plaintiff has suffered. 4 2 On this approach, there is no need for a separate test of remoteness at all within the tort of negligence. If the defendant's duty was not owed in respect of the kind of loss which the plaintiff suffered, the analysis proceeds no further than denying the existence of the duty. If the duty was so owed, then it seems pointless to ask the further question of whether the damage suffered was of a foreseeable kind 4 - unless, of course, one were to think that "foreseeability" or "kinds" of damage were different within the duty and remoteness tests. That would not only be a confusing use of language, but would contravene the underlying principle which was applied in both The Wagon Mound and Caparo v. Dickman: that it is wrong to make a defendant liable for damage which is beyond the risk by reference to which the duty was imposed." This approach within the tort of negligence emphasises that duties are imposed to protect definable interests: it is possible to recover for damage which is physical or economic, caused directly or indirectly by the defendant's act, word or omission, or by third parties. But the 42 This approach is reinforced by the House of Lords' most recent decision relating to damages in contract and tort: South Australia Asset Management Corp. v. York Montague Ltd. [1996] 3 W.L.R. 87, 93 (Banque Bruxelles S.A. v. Eagle Star). 4 Wagon Mound asks the "foreseeability" question directed at the "kind" of damage: [1961) A.C. 388, 426, and it is this basic test which is an unnecessary duplication of the test applied at the point of asking whether a duty of care is owed. Over and above the Wagon Mound test there remain issues such as whether the extent of the damage need be foreseeable (no, according to Hughes v. Lord Advocate [19631 A.C. 837, 845), and whether the unforeseeable injuries of a plaintiff who has an inherent susceptibility ("egg-shell skull") are recoverable (yes, according to Smith v. Leech Brain & Co. Ltd. [1962] 2 Q.B. 405). Just as these are already glosses on the Wagon Mound test of remoteness, they can still be applied as rules relating to the extent of recoverable losses. " This idea was already appreciated at the time of The Wagon Mound itself: Glanville Williams, "The Risk Principle" (1961) 77 L.Q.R Some writers have not sympathised with the Privy Council's concern to ensure congruence between the test of duty and remoteness: e.g. Kidner, "Remoteness of damage: the duty-interest theory and the re-interpretation of the Wagon Mound" (1989) 9 L.S. 1, who separates the questions of (a) who should insure? (duty) and (b) what should he insure for? (remoteness). That article was written before the re-emphasis in Caparo v. Dickman of the categorisation of duties by reference to (inter alia) the foreseeable kinds of loss.

14 The Cambridge Law Journal [19961 reasons for imposing a duty-and so for allowing recovery-for each category have to be looked at in the light of the considerations relevant to that particular category. The question of whether the defendant could reasonably have foreseen the kind of loss suffered by the plaintiff must always be asked. However, we should consider whether, in relation to the various categories of duty, the courts should be asking this "foreseeability" question in precisely the same way. Different Categories of Duty: Hedley Byrne and its Development Wagon Mound was an example of the paradigm case of liability in the tort of negligence: physical damage to property caused by a negligent act. It was therefore natural for the Privy Council not only to see a link between "duty" and "remoteness", but also to see that link in the relatively uncomplicated concept of reasonable foreseeability. 4 5 Since 1961, however, much has changed-and the most significant development came shortly afterwards, in Hedley Byrne v. Heller, 46 which accepted that a duty of care could arise not only in respect of making statements, but also in respect of economic losses which flowed from the failure to exercise such care. Simply to adopt a test of reasonable foreseeability of damage as the test of liability would have been too wide, 47 and the House of Lords sought to devise a narrower test of when a duty of care arises: they limited the duty to cases where the statement was made in the context of a "special relationship"; where the defendant "ought to know that the plaintiff is relying on him to take care"; where he can be said to have "taken upon himself" to take care in making the statement or "undertakes to apply his special skill" for the plaintiff or "assumes a responsibility for" his statement. 48 This approach has been subject to clarification, refinement and development in later cases, and in Caparo v. Dickman it was emphasised that the duty is owed to a person very likely to rely on the statement for a particular purpose. More recently, however, a wider, general principle has been developed-beyond the limited context of statements-in Spring v. Guardian Assurance plc, 49 Henderson v. Merrett Syndicates Ltd. 5 " and White v. Jones. 51 This is not the place for a detailed analysis of these 4 Lord Oliver in Caparo v. Dickman (at pp. 635, 636) also took the view that foreseeability may be the only requirement for the creation of the duty of care in relation to some cases of directly inflicted physical damage. For further refinement, see Marc Rich & Co. AG v. Bishop Rock Marine Co. Ltd [ A.C [1964] A.C ' Lord Reid was careful to point out that words must be treated differently from acts, since the potential consequences are more wide-ranging: he therefore thought that Donoghue v. Stevenson had no direct bearing on the case: ibid., at pp Ibid., at pp. 483, 486, 494, , 505, 514, 525, 529, [1995] 2 A.C o [1995] 2 A.C [1995] 2 A.C. 207.

15 C.L.J. Remoteness of Damage in Contract and Tort 501 cases; but it should be noted that the most recent development has taken Hedley Byrne explicitly beyond its original context of economic loss caused by statements, to economic losses caused by carelessness in performing other services (by both act and omission). The key concept which has been highlighted is the "assumption of responsibility": 2 what places a defendant under a duty of care is that, in performing the service in question, he has-expressly or impliedly 53 - assumed a responsibility towards the plaintiff to take care. But, again, the extent of the duty includes a consideration of the loss sustained: once the defendant assumes a responsibility for particular services, the duty is owed in repect of the losses which (foreseeably) flow from the failure to take care. As Lord Goff said in Henderson: 54 if a person assumes responsibility to another in respect of certain services, there is no reason why he should not be liable in damages for that other in respect of economic loss which flows from the negligent performance of those services. It is consistent with the arguments advanced earlier that the extent of the duty should be defined by reference to the factors which give rise to the duty itself. If the reason for saying that a defendant owed a duty to a particular plaintiff is that he has, on the facts, assumed a responsibility towards him in the provision of certain services, it makes sense to say that the duty extends to the losses of the kinds for which he can be taken to have assumed the responsibility. However, we need to consider what the courts mean by "foreseeability" of loss in this context. The discussion, above, of the remoteness test in contract is relevant here. The House of Lords in The Heron II assumed that the remoteness test in contract should be set at a realistic level, on the basis that the defendant is to bear responsibility for the risks which he undertook in agreeing to the contract; but that tort can be more generous to the plaintiff (in setting a lower threshold for recovery of damage) because the relationship between the parties is inherently different. In the tort of negligence based on the development of Hedley Byrne, however, the relationship between the parties is recognised by the judges as being "equivalent to contract". 55 The 5 Spring v. Guardian, above, at p. 318; Henderson v. Merrett, above, at pp ; White v. Jones, above, at pp. 268, 270, 289. When the question is whether a duty is owed in respect of a statement, it is appropriate to use a test which asks whether the plaintiff was likely to rely on the statement: loss is generally caused by statements not directly but by the plaintiffs own actions in response to the statement: see White v. Jones at p However, when the issue is not a statement but an act (which tends to cause the damage to the plaintiff without the need for the plaintiff's own decision to accept and act upon the defendant's act), the concept of "reliance" by the plaintiff is unhelpful; hence the shift from the plaintiffs "reliance" to the defendant's "assumption of responsibility". Because Hedley Byrne was a case involving statements, either method of expression was equally appropriate. 53 Henderson v. Merrett, above, at p Ibid. 5 Hedley Byrne v. Heller, above, at p. 529 (Lord Devlin), quoted and followed by Lord Goff in Henderson v. Merrett at pp. 179, 181.

16 The Cambridge Law Journal [1996] reason for the duty's existence is not that the law imposes such a responsibility by virtue of the very act being done, or statement being made; it is because the defendant has, or can be taken (objectively 6 ) to have, undertaken a responsibility towards the plaintiff in respect of particular activities. Lord Goff in Henderson v. Merrett 5 7 made it clear that this principle is quite different from the typical case of tort where physical damage is caused by a negligent act of the defendant: in that case there is no "assumption of responsibility". As Lord Reid said in The Heron II,5" the plaintiff in such a case has no way of protecting himself against the risk-he cannot bargain with, for example, the negligent car driver whom he has never met before he runs him overand the careless tortfeasor can therefore not complain at being made to pay for consequences which were "very unusual but nevertheless foreseeable". Clearly, where the existence of the duty is based on an "assumption of responsibility", the dynamics of the relationship between the plaintiff and defendant are different from the case of the negligent car driver. In this context, then, we can expect that, for reasons similar to those used in relation to the remoteness test in contract, a narrower and more realistic form of "foreseeability" of consequences should be used to determine the extent of a defendant's liability. If it is right in the law of contract to draw the line for recoverable damage at the genuinely foreseeable consequences of the breach for the reason that the defendant has agreed to undertake a liability within the scope of that risk, then it ought to be equally right to draw a similar line in those tort cases where the existence of the duty of care depends on a similar notion of a risk assumed voluntarily by the defendant. 59 The courts, in applying the Hedley Byrne line of authority, have required a relatively high degree of likelihood of damage before the defendant can be said to have undertaken a duty in respect of it. In Caparo v. Dickman, for example, the House of Lords made clear that the duty in respect of economic losses flowing from careless statements is only owed when there is a likelihood (and not merely a possibility) 36 Using, according to Lord Goff, the test of objective ascertainment of parties' intentions borrowed from the law of contract: Henderson v. Merrett at p Ibid., at p In Hedley Byrne Lord Devlin also drew a distinction between duties "imposed by law upon certain types of persons or in certain sorts of situations" and (as in Hedley Byrne itself) duties "voluntarily accepted or undertaken": [1964] A.C. 465, [1969] 1 A.C. 350, In White v. Jones at pp Lord Nolan appeared to suggest that all cases of liability in negligence involve an assumption of responsibility: he gave the example of the car driver, who assumes a responsibility towards other road users by taking his car out on the road; the other road users then "rely" on the car driver to discharge his responsibility to take care. This notion of "assumption of responsibility" is however far too wide to be useful in the context we are considering-and is contrary to the view of Lord Goff in Henderson v. Merrett at p The idea of drawing contract and tort together can.occasionally be found: for example, in Canada, BDC Ltd. v. Hofstrand Farms Ltd (1986) 26 D.L.R. (4th) 1, 13-15; this does not however appear to have been pressed to its possible conclusion in later cases.

17 C.L.J. Remoteness of Damage in Contract and Tort of the plaintiff suffering loss by relying on the statement in a particular way. Lord Bridge said:' [T]he defendant giving advice or information was fully aware of the nature of the transaction which the plaintiff had in contemplation, knew that the advice or information would be communicated to him directly or indirectly and knew that it was very likely that the plaintiff would rely on that advice or information in deciding whether or not to engage in the transaction in contemplation. In these circumstances, the defendant could clearly be expected, subject always to the effect of any disclaimer of responsibility, specifically to anticipate that the plaintiff would rely on the advice or information given by the defendant for the very purpose for which he did in the event rely on it... The situation is entirely different where a statement is put into more or less general circulation and may foreseeably be relied on by strangers to the maker of the statement for any one of a variety of different purposes which the maker of the statement has no specific reason to anticipate. This is not to say that the question here of "foreseeability" or "contemplation" will be identical to a case of contract. For example, there is a timing difference. In contract, the issue is what risks the defendant can be said to have undertaken at the time of the contract, rather than the breach-because once the contract is concluded, it is too late to reallocate the risks without re-opening and varying the contract. Therefore, if the defendant is given notice of an unusual potential risk of loss after the contract is concluded but before the breach he is still not affected by it. However, in the case of a duty in tort based on an "assumption of responsibility" there is not such a clear temporal dividing line between the creation of the duty and its breach. The assumption of responsibility in the tort context will typically coincide with the statement, act or omission which is itself relied upon as constituting the breach of the duty. There may be a preexisting assumption of responsibility, for example in the case of a solicitor who assumes a responsibility to his client for the continuing careful performance of his professional activities in conducting the client's affairs. However, this should more properly be characterised as a continuing assumption of responsibility which (in the absence of any contractual term to the contrary) the solicitor could terminate unilaterally. If the approach to remoteness in contract and under the Hedley Byrne cases is similar, this helps to explain why there ought in general to be no problem in the case of an overlap of similar tort and contract duties. Some judges 6 ' have expressed the view that where there is 60 [1990] 2 A.C. 605, See also Lord Oliver at p. 638 and Lord Jauncey at p Even in The Heron H at pp. 411 (Lord Hodson), 413 (Lord Pearce); but also in Parsons v. Uttley Ingham [1978] 1 Q.B. 791, 804 (Lord Denning M.R.), 807 (Scarman L.J.).

18 The Cambridge Law Journal [1996] coincidence of duties in contract and tort, it would be unacceptable to conclude that the plaintiff in pursuing the tort action should, because of a wider remoteness test, be able to recover a wider range of losses than in the contract action. Although it is by no means certain that it would be unacceptable to allow the contract and tort rules to produce different results, 62 in practice in the cases where there is a concident duty in contract and tort based on the same underlying principle-the express or implied undertaking by the defendant of a duty of a particular scope towards the plaintiff-the range of recoverable consequential losses ought to be the same in both. In Esso Petroleum Co. Ltd v. Mardon, for example, the Court of Appeal concluded that there were concurrent duties in contract (an implied term of reasonable care) and tort (a duty of care based on Hedley Byrne) but Lord Denning M.R. made it clear that the recoverable losses were identical. 63 Although no issue of remoteness occurred on the facts, it ought to make no difference whether the plaintiff claimed under the tort duty or the contract duty, since the tort duty arose from the same factors as the contractual duty of care: the holding-out by Esso during the pre-contractual negotiations that they were skilled and would exercise care in making estimates of the potential throughput of the petrol station which Mr. Mardon was proposing to lease. The argument for remoteness tests which draw the same line in cases of overlap between contract and tort only holds, however, where the basis and content of the contract and tort duties are identical. In a case where a plaintiff has claims in both tort and contract against a defendant, but the claims are in respect of different kinds of obligation (such as an absolute obligation in contract but a duty of reasonable care in tort) or where the sources of the obligations are not coincident (if, for example, the tort duty arises out of statements or acts which are not themselves covered at all by the terms (express or implied) of the contract), there should be no expectation that the measure of recovery in both claims would be identical: each area of law should apply its own policies which would not then necessarily coincide. COMPARISON OF THE POLICIES REFLECTED IN THE REMOTENESS TESTS IN CONTRACT AND NEGLIGENCE We can now reconsider the operation of the policies discussed at the beginning of this article. In both contract and negligence, the courts appear to take the view that the range of losses for which the defendant is responsible is linked to the underlying policy pursuant to which the 62 And, indeed, the House of Lords said in Henderson v. Merrett, in the context of limitation of actions, that a plaintiff should be entitled to pursue whichever claim is the more advantageous for him. 63 [1976] Q.B. 801, 820.