Manchester Building Society v Grant Thornton UK LLP

This appeal is concerned with the application of the concept of scope of duty in the tort of negligence, as illustrated by the decision of the House of Lords in Banque Bruxelles Lambert SA v Eagle Star Insurance Co Ltd; South Australia Asset Management Corpn v York Montague Ltd [1997] AC 191 (“SAAMCO”) in relation to recovery of damages for economic loss. The context is professional advice given by expert accountants. The appeal was heard by the same expanded constitution of the court which heard the appeal in Khan v Meadows [2021] UKSC 21, which is concerned with the same issue in the context of professional advice given by a medical expert. The reason the appeals were heard by the same constitution of the court was to provide general guidance regarding the proper approach to determining the scope of duty and the extent of liability of professional advisers in the tort of negligence. It is therefore desirable that the judgments in the two appeals should be read together as reflecting and supporting a coherent underlying approach. The present judgment should be read with our judgment in Khan v Meadows.

Accountancy advice is usually given pursuant to a contract, as was the valuation advice in SAAMCO and the legal advice considered in the other leading judgment in this area, Hughes-Holland v BPE Solicitors [2017] UKSC 21; [2018] AC 599 (“Hughes-Holland”). In such cases, there is a parallel duty of care in tort and in contract. The extent of the responsibility assumed by the professional adviser, and the extent of their liability if they fail to act with reasonable care, is the same in tort and in contract. Medical advice may also be given pursuant to a contract, in the private medical sector. There too there is a parallel duty of care in tort and in contract, and the extent of the responsibility assumed by the professional adviser and the extent of their liability will again be the same. In what follows, for ease of exposition we will focus on the scope of the duty of care in tort. The scope of the parallel duty of care in contract depends on the same factors.

In the present appeal we have had the benefit of reading the judgment of Lord Leggatt. We agree with much of it. In particular, we find his explanation of the valuer cases illuminating. We have reservations about his explanation of the auditor negligence case involving a payment of dividends (paras 130-131), as it is not obvious to us why recovery of damages should be limited to a payment out of capital which would not have been made but for the negligent advice but would not be capable of covering a payment out of retained profits which would not have been made but for that advice, where the sum paid would otherwise have been retained by the company as available working capital. But it is not necessary to discuss that issue further in this judgment. We are grateful for his account of the facts of the case, albeit there are certain features of the facts found by the trial judge, Teare J, which we think require greater emphasis. We address those below. We agree with the outcome of the appeal which Lord Leggatt proposes and with which Lord Burrows also agrees. But we find ourselves, with respect, unable to support Lord Leggatt’s approach to SAAMCO and the question of the scope of the duty of the accountants, Grant Thornton, in this case. Our approach in the present case is closer to, but not fully aligned with, that of Lord Burrows. Given the importance of the issues, we think we should explain our own view. The divergence of opinion about SAAMCO and the scope of duty principle at this level serves to emphasise the importance of seeking to arrive at an authoritative view after debate within the court.

In summary, our view is that (i) the scope of duty question should be located within a general conceptual framework in the law of the tort of negligence; (ii) the scope of the duty of care assumed by a professional adviser is governed by the purpose of the duty, judged on an objective basis by reference to the purpose for which the advice is being given (in the context of this judgment, we use the expression “purpose of the duty” in this sense); (iii) in line with the judgment of Lord Sumption in Hughes-Holland at paras 39-44, the distinction between “advice” cases and “information” cases drawn by Lord Hoffmann in his speech in SAAMCO should not be treated as a rigid straitjacket; and, following on from this, (iv) counterfactual analysis of the kind proposed by Lord Hoffmann in SAAMCO should be regarded only as a tool to cross-check the result given pursuant to analysis of the purpose of the duty at (ii), but one which is subordinate to that analysis and which should not supplant or subsume it. The points which we make below in relation to the facts of the case as found by the judge reflect our view regarding the proper approach to be adopted.

In our view the scope of duty principle can more readily be understood without placing the emphasis which Lord Leggatt does on causation and the counterfactual test; the focus on causation, which gave rise to the debate discussed in Hughes-Holland at paras 37 and 38, distracts attention from the primary task of identifying the scope of the defendant’s duty. Subject to what we say below about the facts, we are in broad agreement with what Lord Burrows says about those matters. However, we differ from Lord Burrows in our understanding of the location of the scope of duty issue in the scheme of the law of tort and in thinking that the focus for the analysis of that issue should be on the purpose of the duty without involving reference back to policy. The policy decision has already been made that the proper approach to the scope of duty issue is to derive it from the purpose of the duty. In our opinion it is unnecessary to reintroduce a policy-based analysis and to do so would create the risk of uncertainty: see our judgment in Khan v Meadows, para 59. A focus on the purpose of the duty is, in our view, both more principled and more in line with authority.

Lord Sumption explained in Hughes-Holland at paras 20-29 that the idea of limiting the damages recoverable in the tort of negligence to those falling within the scope of the duty of care assumed by the defendant long pre-dated the decision in SAAMCO. As we say in Khan v Meadows, para 28, it is helpful to analyse the place of the scope of duty principle in the tort of negligence in the following way. When a claimant seeks damages from a defendant in the tort of negligence, a series of questions arise:

Is the harm (loss, injury and damage) which is the subject matter of the claim actionable in negligence? (the actionability question)

What are the risks of harm to the claimant against which the law imposes on the defendant a duty to take care? (the scope of duty question)

Did the defendant breach his or her duty by his or her act or omission? (the breach question)

Is the loss for which the claimant seeks damages the consequence of the defendant’s act or omission? (the factual causation question)

Is there a sufficient nexus between a particular element of the harm for which the claimant seeks damages and the subject matter of the defendant’s duty of care as analysed at stage 2 above? (the duty nexus question)

Is a particular element of the harm for which the claimant seeks damages irrecoverable because it is too remote, or because there is a different effective cause (including novus actus interveniens) in relation to it or because the claimant has mitigated his or her loss or has failed to avoid loss which he or she could reasonably have been expected to avoid? (the legal responsibility question)

The first question is a threshold question and asks whether the matter about which the claimant complains is actionable. A claimant may have a cause of action in negligence to recover damages for physical injury, psychiatric injury, damage to property and economic loss, but not all complaints are actionable in negligence. Personal upset or annoyance and diminished enjoyment of a person’s property resulting from noises or smells are not actionable in negligence. A defendant may act carelessly without incurring liability to a claimant in the absence of actionable loss.

The second question addresses the scope of a defendant’s duty and is the central question in this appeal. The fact that the defendant owes the claimant a duty to take reasonable care in carrying out its (the defendant’s) activities does not mean that the duty extends to every kind of harm which might be suffered by the claimant as a result of the breach of that duty. In Spartan Steel Ltd & Alloys v Martin & Co (Contractors) Ltd [1973] QB 27, for example, the duty of care owed by workmen not to cut off electrical power to the claimant’s factory was imposed in order to protect the claimant from suffering damage to its property, so the claimant could only sue for damages to compensate it for property damage it had suffered as a result of the breach of the duty of care, and not for damages to compensate it for the distinct loss of business it had suffered as a result of the loss of power. Similarly, in Caparo Industries plc v Dickman [1990] 2 AC 605 (“Caparo”) it was recognised that although the auditor of a company’s accounts owes a duty of care to shareholders in the company for some purposes, breach of that duty does not mean that a shareholder can claim damages for loss flowing from its reliance on the audited accounts to make investment decisions (p 627 per Lord Bridge of Harwich; pp 651-653 per Lord Oliver of Aylmerton; pp 660-662 per Lord Jauncey of Tullichettle). As Brennan J stated in Sutherland Shire Council v Heyman (1985) 157 CLR 424, at 487:

The discussion in SAAMCO regarding the scope of the duty of care is relevant to the second question. SAAMCO was concerned with a single type of loss, namely pure economic loss. In Platform Home Loans Ltd v Oyston Shipways Ltd [2000] 2 AC 190, Lord Hobhouse of Woodborough made the acute observation at p 209G that Lord Hoffmann’s development of scope of duty reasoning in SAAMCO was that “instead of applying it to kinds or categories of damage,” he “applied it to the quantification of damage” (emphasis in the original). Some confusion has arisen from references in the cases to “the SAAMCO principle”, whereas on proper analysis SAAMCO is not a distinct principle but rather is an illustration in a particular context of the scope of duty principle.

It is a basic element of a cause of action in negligence that the claimant can allege that he has suffered loss falling within the scope of a duty of care owed to him by the defendant. That is why in SAAMCO Lord Hoffmann said that the Court of Appeal in that case, by moving directly to ask what damages should be awarded to put the claimant in the position he would have been in had the breach not occurred, had started in “the wrong place” (p 211, quoted in Hughes-Holland at para 27). It had failed to ask whether and to what extent the loss for which damages were claimed was within the scope of the duty of care. As Lord Hoffmann said in the same passage, “[a] correct description of the loss for which the valuer is liable must precede any consideration of the measure of damages. For this purpose it is better to begin at the beginning and consider the lender’s cause of action.” Consideration of the scope of the duty of care owed by the defendant valuers led to the conclusion that they were not responsible in law for the full extent of the loss suffered by the claimant banks.

For the same reason, the burden of proof lies on the claimant to show that the loss for which he claims damages lies within the scope of the duty of care owed to him by the defendant: see SAAMCO, p 220 per Lord Hoffmann, and Hughes-Holland, para 53 per Lord Sumption. That is the essence of the claimant’s cause of action in tort. This is also the point made, perhaps slightly cryptically, by Lord Sumption in Hughes-Holland at para 38 when he says that the question posed by the scope of duty principle as illustrated by SAAMCO is not one of causation in the conventional sense of identifying the consequences which flow from a breach of duty, but “rather whether the loss flowed from the right thing, ie from the particular feature of the defendant’s conduct which made it wrongful. That turns on an analysis of what did make it wrongful.”

In some cases, a claim may be answered at stage 2 without the need to address the questions of breach and factual causation. However, in cases where the scope of duty question is relevant to the extent of loss of a particular kind, as in SAAMCO and Hughes-Holland, it is generally more appropriate to examine this after first ascertaining on a simple “but for” basis what is the extent of the loss which has flowed from the alleged breach of duty. Proceeding in this way means that one identifies the losses which are in fact in issue so that it is possible to focus with greater precision on the extent to which they fall within the scope of the duty of care owed by the defendant. This was the approach adopted in the valuer negligence cases which followed SAAMCO. As Lord Nicholls explained in Nykredit Mortgage Bank plc v Edward Erdman Group Ltd (formerly Edward Erdman (an unlimited company) (No 2) [1997] 1 WLR 1627, 1631, one begins by identifying what he called “the basic measure” of the claimant’s loss and what Lord Hobhouse in Platform Home Loans described as “the basic loss” which the claimant has suffered (ie the loss which can be identified as flowing from the alleged breach of duty as a matter of “but for” factual causation), and then examines the extent to which that loss falls within the scope of the duty assumed by the valuer (see also Hughes-Holland, para 31, and our judgment in Khan v Meadows, para 52). This is the reason why in this sort of case it is appropriate to ask the duty nexus question at stage 5. But it should be recognised that this is simply a practical approach to working out the implications of the scope of duty concept which arises, in principle, earlier in the analysis, at stage 2.

In our respectful opinion, the scope of duty question can and should be approached in a more straightforward way than is suggested by Lord Leggatt. In our view, the scope of the duty of care assumed by a professional adviser is governed by the purpose of the duty, judged on an objective basis by reference to the reason why the advice is being given (and, as is often the position, including in the present case, paid for). Lord Hoffmann was explicit about this in SAAMCO at p 212:

The other leading authorities confirm that this is the proper approach. All the speeches in Caparo emphasised that the scope of the auditor’s duty of care in that case was to be derived from an examination of the purpose served by the duty to audit the accounts of the company. This observation by Lord Roskill is representative (p 629B): “… before the existence and scope of any liability can be determined, it is necessary first to determine for what purposes and in what circumstances the information in question is to be given.” See also, in particular, p 627C-D per Lord Bridge; p 652 (“In seeking to ascertain whether there should be imposed on the adviser a duty to avoid the occurrence of the kind of damage which the advisee claims to have suffered … [o]ne must … ask, in what capacity was his interest to be served and from what was he intended to be protected?”; “Before it can be concluded that the duty is imposed to protect the recipient against harm which he suffers by reason of the particular use that he chooses to make of the information which he receives, one must … first ascertain the purpose for which the information is required to be given”) and p 654 per Lord Oliver; and pp 655D-E and 660D-F per Lord Jauncey (auditors are aware that the audited accounts will be seen and relied on by shareholders, but “that does not answer the fundamental question of the purpose, and hence the very transactions, for which the annual accounts of a company are prepared and distributed to its members”).

Similarly, in Aneco Reinsurance Underwriting Ltd (in liquidation) v Johnson & Higgins Ltd [2001] UKHL 51; [2002] 1 Lloyd’s Rep 157 (“Aneco”) (discussed in Hughes-Holland at para 43) the House of Lords addressed the issue of the extent of the liability of the defendant broker by analysing the purpose of the task which he undertook. The claimant reinsurer proposed to reinsure a book of excess of loss business, but only if he could retrocede part of the book into the market, and instructed the defendant to place the retrocession. The defendant reported that he had placed it, but he had failed to represent the risk fairly, with the result that after the claimant had written the reinsurance his retrocession was avoided. The appellate committee was divided as to the relevant factual analysis and the result, but agreed as to the approach to be adopted. The majority held that the claimant was entitled to recover damages for its full loss flowing from the writing of the reinsurance because, on their view of the facts, one part of the purpose of the duty assumed by the defendant was to advise the claimant whether there was any market at all for the book of business to be written by him; the defendant should have advised that there was none; and if so advised the claimant would have declined to write any part of the business: see para 17 per Lord Lloyd of Berwick (the difference as to the scope of the duty did not depend on calling one aspect of the advice given “information” and the other “advice”, according to the suggestion by Lord Hoffmann in SAAMCO; rather, “[i]t depends on a difference of substance … on the scope of the advice which the brokers undertook to give”) and paras 23 and 40 per Lord Steyn. Lord Millett, dissenting on the facts, also proceeded by analysing the purpose of the duty assumed by the defendant: at para 64 he cited Lord Roskill’s speech in Caparo at p 629, quoted above; at paras 95-100 he analysed the purpose for which the defendant was engaged to give advice; and at para 110 he concluded that the full extent of the loss claimed did not fall within the scope of the defendant’s duty because learning about the general state of the reinsurance market for this kind of business was “not the purpose for which Aneco wanted to know that cover was available”.

In Hughes-Holland, Lord Sumption analysed the cases in this way and again emphasised the importance of the purpose of the duty: see para 23 (Caparo), para 28 (SAAMCO), para 43 (Aneco) and paras 54-55 (explaining the outcome in Hughes-Holland itself, on the basis that the defendant solicitors did not assume responsibility for their client’s decision to lend money which was then lost, since “[t]heir instructions were to draw up the facility agreement and the charge, nothing more”). Although Lord Sumption uses the concept of assumption of responsibility in his judgment (see, in particular, paras 44 and 54), it is clear that this is found to arise where the defendant adviser has taken on responsibility for a particular task having a particular purpose.

Therefore, in our view, in the case of negligent advice given by a professional adviser one looks to see what risk the duty was supposed to guard against and then looks to see whether the loss suffered represented the fruition of that risk. This is the point of the mountaineer’s knee example given by Lord Hoffmann in SAAMCO at p 213.

The distinction drawn by Lord Hoffmann in SAAMCO between “advice” cases and “information” cases has not proved to be satisfactory. Put shortly, as explained by Lord Sumption in Hughes-Holland at paras 39-44, the distinction is too rigid and, as such, it is liable to mislead. In reality, as Lord Sumption emphasises at para 44, the whole varied range of cases constitutes a spectrum. At one extreme will be pure “advice” cases, in which on analysis the adviser has assumed responsibility for every aspect of a transaction in prospect for his client. At another extreme will be cases where the professional adviser contributes only a small part of the material on which the client relies in deciding how to act. In some cases (such as those involving valuers) it is readily possible to say that the purpose of the advice given is limited and that the adviser has assumed responsibility under a duty the scope of which is delimited by that purpose, which Lord Hoffmann called an “information” case. However, Lord Sumption observed (para 44), “[b]etween these extremes, every case is likely to depend on the range of matters for which the defendant assumed responsibility and no more exact rule can be stated”.

In our view, for the purposes of accurate analysis, rather than starting with the distinction between “advice” and “information” cases and trying to shoe-horn a particular case into one or other of these categories, the focus should be on identifying the purpose to be served by the duty of care assumed by the defendant: see section (ii) above. Ascribing a case to one or other of these categories seems to us to be a conclusion to be drawn as a result of examination of that prior question.

This also corresponds with Lord Sumption’s explanation at paras 40 and 41 of what is involved in an “advice” case and an “information” case, respectively. In an “advice” case, the adviser’s duty “is to consider all relevant matters and not only specific factors” (and what counts as a relevant matter for the adviser is determined by the purpose for which he has agreed to give advice: see para 44). Where the adviser is responsible for guiding the whole decision-making process, the adviser’s responsibility extends to the decision. In that circumstance, as Lord Sumption explains (para 40), “[if] the adviser has negligently assessed risk A, the result is that the overall riskiness of the transaction has been understated. If the client would not have entered into the transaction on a careful assessment of its overall merits, the fact that the loss may have resulted from risks B, C or D should not matter”.

By contrast, in an “information” case (Hughes-Holland, para 41), the adviser contributes a limited part of the material to be relied on, “but the process of identifying the other relevant considerations and the overall assessment of the commercial merits of the transaction are exclusively matters for the client” (emphasis added), and in such a case “the defendant’s legal responsibility does not extend to the decision itself”; the result then is that the defendant is “liable only for the financial consequences of [the information] being wrong and not for the financial consequences of the claimant entering into the transaction so far as these are greater”.

We welcome Lord Leggatt’s proposal (para 92) to dispense with the descriptions “information” and “advice” to be applied as terms of art in this area. As Lord Sumption points out in Hughes-Holland, para 39, both “advice” and “information” cases involve the giving of advice. For the reasons we give, we think it is important to link the focus of analysis of the scope of duty question and the duty nexus question back to the purpose of the duty of care assumed in the case in hand.

Related to the issues examined in sections (i) to (iii) above is the use of counterfactual analysis as set out by Lord Hoffmann in SAAMCO. Lord Hoffmann proposed a form of counterfactual analysis as a way to assist in identifying the extent of the loss suffered by the claimant which falls within the scope of the defendant’s duty, by asking in an “information” case whether the claimant’s actions would have resulted in the same loss if the advice given by the defendant had been correct. This procedure generates a limit to the damages recoverable which has been called the SAAMCO “cap”. As Lord Sumption said in Hughes-Holland, para 45, this is “simply a tool for giving effect to the distinction between (i) loss flowing from the fact that as a result of the defendant’s negligence the information was wrong [ie the loss falling within the scope of the defendant’s duty] and (ii) loss flowing from the decision to enter into the transaction at all [ie by application of a simple “but for” test]”. As so explained, it is clear that the use and, in particular, the correct framing of the counterfactual scenario follows from the prior question, which is, what purpose was the duty of care assumed by the defendant supposed to serve? In that regard, we agree with Lord Burrows (paras 195-203) that the counterfactual test may be regarded as a useful cross-check in most cases, but that it should not be regarded as replacing the decision that needs to be made as to the scope of the duty of care (albeit he describes that as a policy decision, whereas we think it reflects more fundamental issues of principle: see section (ii) above).

In SAAMCO the House of Lords had to explain why the Court of Appeal in that case had erred by simply applying the general tests of foreseeability and remoteness of loss, and the counterfactual analysis was deployed by Lord Hoffmann as a way of giving emphasis to the importance of the scope of duty principle which he applied in that case. It helped to show why the valuers could not be taken to have assumed responsibility for the whole loss flowing from the fall in the market. But it would have been sufficient to arrive at that conclusion (and to determine the amount of recoverable damages) to ask what was the purpose of the valuer’s duty to advise: it was to allow the lender to determine at market values current at the time of the advice the amount of security which it would take.

Also, linking the use of the counterfactual analysis to “information” cases in the “advice” and “information” framework is unhelpful, because of the problems associated with that framework: see section (iii) above. By contrast, examination of the purpose of the duty provides an appropriate and refined basis for identifying, out of what may be a wide range of factors which contribute to the claimant’s loss, the factors for which defendant is responsible.

Another problem associated with counterfactual analysis of this kind is the danger of manipulation, in argument, of the parameters of the counterfactual world. Lord Leggatt points out (paras 128-132) that the counterfactual test can yield the right result if it is properly applied. However, the more one moves from the comparatively straightforward type of situation in the valuer cases, as illustrated by SAAMCO, the greater scope there may be for abstruse and highly debatable arguments to be deployed about how the counterfactual world should be conceived. One has to take care, therefore, not to allow the counterfactual analysis to drive the outcome in a case. To do so would create a risk of litigation by way of contest between elaborately constructed worlds advanced by each side, which would become increasingly untethered from reality the further one moves from the relatively simple valuer case addressed in SAAMCO. There was an element of this in the present appeal, as Mr Salzedo QC for Grant Thornton sought to persuade us that the counterfactual world in this case should be constructed in such a way as to show that Manchester Building Society (“the society”) would have suffered the same loss if Grant Thornton’s advice had been correct and the society responded in kind with elaborations of its own. This in part explains why an aspect of the society’s submissions took the form which Lord Leggatt criticises at para 151 et seq, even though its pleaded case was a straightforward one to the effect that Grant Thornton were aware, when they advised in 2006 and thereafter, of the commercial significance of hedge accounting for the society in terms of its impact on its regulatory capital position and, had non-negligent advice been given, the society would not have engaged in the business of matching swaps and mortgages at all and would not have been exposed to the loss which it eventually suffered when it had to unwind that business to protect and so far as possible restore its capital position. Lord Leggatt engages in a sophisticated analysis to answer the elaborate variants of the submissions advanced by the parties (para 143 et seq), but the fact is that a distinguished constitution of the Court of Appeal fell into error because of them. Again, it seems to us that the better approach is to focus more directly on the purpose for which the defendant gave the advice in question. There is no need to apply a counterfactual test to arrive at the correct conclusion and it has the potential to confuse rather than assist the correct analysis.

The points which we make in this judgment are interrelated. Identifying the scope of the duty of care by reference to its purpose is a reasonably determinate test, applicable in principle from the outset of the parties’ relationship. It seems to us that a focus on this criterion is a surer and simpler guide than a causation-based analysis as proposed by Lord Leggatt. It is fair to say that the two modes of analysis may often lead to the same outcome, but problems arise where it is unclear whether they do or not. A choice then has to be made, and in our view it should be in favour of clear adoption of the purpose of the duty of care as the relevant test. Analysis using the counterfactual “tool” as deployed in SAAMCO was designed to assist with looking at the scope of duty question from a causation-based perspective. Therefore, once it is accepted that the scope of duty inquiry turns on identifying the purpose of the duty, it can readily be seen that a SAAMCO-type counterfactual analysis is just a cross-check, rather than the foundation of the relevant analysis. By contrast, if emphasis is given to a causation-based analysis of the scope of duty question and the related duty nexus question, then SAAMCO-type counterfactual analysis moves centre stage and appears to assume greater significance than it should do.

The present case has some unusual features which differentiate it from the type of valuer case illustrated by SAAMCO. Grant Thornton advised the society in circumstances where the management of the society had made their own assessment about the nature of the commercial markets for lifetime mortgages and for swaps and had made their own judgment that a business model matching swaps and mortgages would be commercially attractive. This is not a case in which Grant Thornton was asked to give advice about these matters. Also, the management of the society understood the true underlying financial position of the society. They appreciated that the mark-to-market value of swaps was subject to constant variation and that the society would have to make payments to swaps counterparties reflecting varying interest rates. They had made the commercial decision for themselves that the interest payments in relation to the swaps would be matched by those to be received by the society under the mortgages, which would protect the society over the life of the swaps as interest rates happened to change. Grant Thornton was not asked to provide commercial advice about these matters.

However, the society had an interest in the accounting treatment of the swaps and the mortgages from a distinct commercial perspective. As a lending institution, the society was subject to regulation. At the material time, the regulatory authority was the Financial Services Authority (“the FSA”). Under the regulatory regime, the society was required to maintain a substantial level of capital to ensure its continuing viability should it come under stress, which is referred to as “regulatory capital”. If it failed to do so, the FSA could take steps to close its operations. Also, the more the society’s financial activities were subject to volatility, the higher the level of regulatory capital it was required to have in place as a safeguard. The tool which the FSA used to monitor volatility and the level of regulatory capital the society was required to maintain was its accounts.

In 2005 the society changed the format for the preparation of its accounts from one set of accounting standards, under which swaps were not included on its balance sheet, to the International Financial Reporting Standards (“IFRS”), under which swaps did have to be brought onto its balance sheet. In the context of the IFRS, Lord Leggatt has explained the significance of application of the “hedge accounting” convention. Application of that convention in the society’s accounts would have the effect that swaps were matched with the society’s mortgage book, thereby greatly reducing the appearance of volatility in the society’s profits and greatly reducing the level of capital which it would be required to maintain to meet regulatory requirements.

Teare J made findings that Grant Thornton understood from their discussions with the society’s management the regulatory capital issue and the importance for the society of being able to use hedge accounting as a “key tool in avoiding the profit volatility caused by recognising … hedging instruments at fair value” (para 162). He also found that in discussions between the society and Grant Thornton in late 2005 and early 2006 the society was looking to Grant Thornton for advice whether it was entitled to use hedge accounting in drawing up its accounts and that it needed that advice in order to make a commercial decision whether to enter into swaps to be matched against mortgages, in other words to begin to carry into effect its proposal for a new business model involving matching swaps and lifetime mortgages. He further found that Grant Thornton were also asked to advise on whether the practice which the society was proposing to adopt of substituting different lifetime mortgages in the future against long term swaps would be permitted under the hedge accounting rules and that they confirmed by an email dated 11 April 2006 that this would be permitted and on that basis approved the society’s proposal to use hedge accounting in drawing up its accounts. The significance of this was that in regulatory capital terms the society would be able to afford to implement its business model.

In reliance on that advice, the society decided to pursue the matched swaps and mortgages business model. It decided not to unwind its two existing swaps and it entered into further swaps. It did so not in order to trade in swaps or to treat them as a speculative investment, but (as Grant Thornton knew) with a view to holding them to maturity, matching them against its book of lifetime mortgages.

Grant Thornton’s advice was repeated thereafter each year it signed an audit opinion stating that the society’s annual accounts, drawn up on the basis of the hedge accounting policy, gave a true and fair view of its financial position. Their original advice was related to this later advice, since the original advice was given to inform the society that it would be appropriate to draw up its accounts using hedge accounting and that Grant Thornton would be willing to sign audit opinions in the future to certify that the society’s accounts so prepared did give a true and fair view. However, the original advice given in 2006 was particularly significant, since it was on the basis of that advice that the society entered into new swaps (and decided not to unwind the existing swaps) and the disruption of financial markets in the financial crash of 2008 followed soon afterwards, resulting in a sharp fall in interest rates and exposing the society to the risk of significant financial loss if it were to be forced to break the swaps.

In the circumstances in which Grant Thornton gave its advice, the purpose of the advice was clear. They advised that the society could employ hedge accounting in order to reduce the volatility on its balance sheet and keep its regulatory capital at a level it could afford in relation to swaps to be held to term on the basis that they were to be matched against mortgages. In other words, the society looked to Grant Thornton for technical accounting advice whether it could use hedge accounting in order to implement its proposed business model within the constraints arising by virtue of the regulatory environment, and Grant Thornton advised that it could. That advice was negligent. It had the effect that the society adopted the business model, entered into further swap transactions and was exposed to the risk of loss from having to break the swaps, when it was realised that hedge accounting could not in fact be used and the society was exposed to the regulatory capital demands which the use of hedge accounting was supposed to avoid. That was a risk which Grant Thornton’s advice was supposed to allow the society to assess, and which their negligence caused the society to fail to understand.

In our view, in the light of the findings he had made, Teare J was essentially correct in his summary at paras 172-173 in so far as he said:

However, the judge did not draw the conclusion that Grant Thornton were liable for the losses suffered by the society in being compelled to break the swaps once the true accounting position was appreciated. It is our view, with respect, that he should have done. As he said at para 149, “[i]t was the volatility of the balance sheet which led to the swaps being closed in 2013”, and hence to the loss suffered by the society. On the judge’s findings the society had suffered a loss which fell within the scope of the duty of care assumed by Grant Thornton, having regard to the purpose for which they gave their advice about the use of hedge accounting.

The judge described the society’s argument as “cogent”, but then gave a series of reasons why he did not accept it. We agree with Lord Leggatt (paras 169-172) that those reasons cannot be supported.

In one respect our analysis is similar to that set out by Lord Leggatt at paras 166-168. He relies on Grant Thornton’s misrepresentation about the existence of an “effective hedging relationship” between the swaps and the mortgages as the reason why the society’s loss fell within the scope of its duty of care. We agree that this is critical. But for the purposes of analysing whether the loss suffered by the society fell within the scope of the duty of care owed by Grant Thornton we think it is important to have regard to the commercial reason, as appreciated by Grant Thornton, why advice about this was being sought and why this was fundamental to the society’s decision to engage in the business of matching swaps and mortgages. That reason was the impact of hedge accounting on the society’s regulatory capital position. Use of hedge accounting allowed the society to make the assessment that, in terms of the constraints imposed on it by the regulatory capital requirements to which it was subject, it had the capacity to proceed with that business whereas otherwise it did not. In our opinion, reference to the reason the advice was sought and given is important, because that is the foundation for the conclusion that the purpose of the advice was to deal with the issue of hedge accounting in the context of its implications for the society’s regulatory capital. It is not in dispute that the loss in issue formed part of the society’s “basic loss” flowing from Grant Thornton’s negligent advice. Examination of the purpose for which that advice was given shows that the loss fell within the scope of their duty of care. Having regard to that purpose, we consider that Grant Thornton in 2006 in effect informed the society that hedge accounting could enable it to have sufficient capital resources to carry on the business of matching swaps and mortgages, when in reality it did not. In our opinion, this is analogous to a dividend payment case, where an auditor negligently advises a company that it has capital resources at a level which would permit payment of a dividend when in fact it does not.

We also agree with what Lord Leggatt says about legal causation: paras 173-174. Further, like him, we consider that the judge was entitled to make the assessment that the society’s damages should be reduced by 50% on the basis of its contributory negligence. The contribution by the society to its own loss arose from the mismatching of mortgages and swaps in what was an overly ambitious application of the business model by the society’s management.

On this appeal (and that in Khan v Meadows [2021] UKSC 21, heard by the same panel of seven Justices) this court is once more asked to consider questions about the extent of a professional person’s liability for loss caused by giving negligent advice. A firm of accountants incorrectly and negligently advised its client, a building society, that the society’s accounts could be prepared using a method known as “hedge accounting” and that the accounts prepared using that method gave a true and fair view of the society’s financial position. In reliance on that advice, the society carried on a strategy of entering into long-term interest rate swap contracts as a hedge against the cost of borrowing money to fund mortgage lending. The misstated accounts served to hide volatility in the society’s capital position and what became a severe mismatch between the negative value of the swaps and the value of the mortgage loans which the swaps were supposed to hedge. When after seven years the accountants realised their error, the society had to restate its accounts to show substantially reduced net assets and insufficient regulatory capital. To extricate itself from this predicament, the society closed out the swaps at a cost of over £32m.

The issue on this appeal is whether the society can recover this cost as damages from the accountants (reduced by 50% for the society’s contributory negligence). The trial judge and the Court of Appeal held that it cannot. Their reasoning, although different, was in each case based on their understanding of the principle illustrated by the leading case of South Australia Asset Management Corpn v York Montague Ltd [1997] AC 191 (“SAAMCO”). This principle has proved difficult to formulate as well as difficult to apply, but is generally expressed by saying that a professional adviser is only liable for losses which are “within the scope” of the adviser’s duty of care. The scope of duty principle was recently clarified by this court in Hughes-Holland v BPE Solicitors [2017] UKSC 21; [2018] AC 599. However, its application continues to give rise to difficulties, as the present case has shown.

For the reasons explained in this judgment, I consider that, on a correct appreciation of the principle, it does not limit the extent of the accountants’ liability in damages in this case.

The claimant (and appellant) is a small mutual building society, which I will refer to as “the society”. From 1997 until 2012, the society’s accounts were audited by the defendant (and respondent to this appeal), Grant Thornton UK LLP, a large firm of accountants.

Between 2004 and 2010 the society purchased and issued lifetime mortgages. These were mortgage loans designed to release the equity in the borrower’s home. Interest on the loan was charged at a fixed rate but neither the interest nor the capital sum was repayable until the borrower died, moved out of the property or chose to repay the loan and redeem the mortgage. Until that time, which was necessarily uncertain, the interest on the loan was compounded. The borrowers under these lifetime mortgages were homeowners (over the age of 50) in the United Kingdom and Spain. Between 2004 and 2008 the society acquired UK lifetime mortgages with a total value of £68m; and between 2008 and 2010 the society issued Spanish lifetime mortgages with a total value of over £40m.

The mortgage loans were funded by borrowing at variable rates of interest. In order to protect itself against the risk that the variable cost of borrowing would exceed the fixed rate of interest receivable on the mortgage loans, the society entered into interest rate swap contracts. Under these swaps, the society agreed to pay a fixed rate of interest on a notional sum in return for the counterparty’s agreement to pay a variable rate of interest on the same sum. The interest payments were periodically netted off against each other with the balance payable by one party to the other (depending on whether the fixed or variable interest rate was higher during the period). The intention was that the variable rate of interest payable by the swap counterparties should match the variable rate payable by the society to borrow money, while the fixed rate payable by the society under the swaps was less than the fixed rate of interest receivable under the lifetime mortgages. In this way, when the lifetime mortgages were eventually redeemed, the society would be guaranteed to make a profit.

For the hedge to be effective, the value and duration of the swaps needed to match the value and duration of the mortgages.

Between May 2006 and February 2012, the society entered into 14 additional interest rate swaps (leaving aside one which was replaced by another) to hedge the UK lifetime mortgages. The total notional value of these swaps was £74.2m and most were for a term of 50 years. Between July 2008 and May 2012, the society also entered into 14 interest rate swaps to hedge the Spanish lifetime mortgages, with a total notional value of €57m. These swaps were for terms of between five and 30 years.

The “mark-to-market” (“MTM”) value of a swap is the price (or estimated price) for which it can be traded in the market at a given date. This price is calculated by estimating the value of all the future payments to be made over the remaining term of the swap and discounting these payments to a net present value. At the very beginning and end of the term, the MTM value will be zero. It will be zero at the beginning because the swap is priced to reflect the market rate of interest (for the term of the swap) at that time. It is zero at the end because there are no further periodic payments to make. Between those dates the MTM value of the swap will fluctuate according to the market’s forecast of future interest rates over the remaining term of the swap. The MTM value is not a sum that either party will actually pay or receive, unless the swap is traded or terminated prematurely. However, when the MTM value of a swap is negative, the party “out of the money” may be obliged to provide collateral in the amount that would be payable to the counterparty on early termination.

From 2005 onwards, the society was required to prepare its accounts in accordance with the International Financial Reporting Standards, which require swaps to be accounted for on the balance sheet at their fair value. The fair value of a swap is its MTM value. The mortgage loans, on the other hand, were accounted for at their amortised cost (or book value). A consequence of accounting for the swaps at fair value was that the value shown on the balance sheet would reflect movements in interest rates. The society’s reported financial position would accordingly become volatile. This volatility would in turn increase the amount of capital needed to satisfy regulatory requirements.

However, such volatility could be mitigated if (and only if) the society was able to use “hedge accounting”. Where hedge accounting is permitted, the carrying value of the hedged item (here, the lifetime mortgages) can be adjusted to offset changes in the fair value of the hedging instrument (here, the swaps), thus reducing accounting volatility.

In April 2006, Grant Thornton advised the society that it could apply the hedge accounting rules under International Accounting Standard 39 (“IAS 39”) to the lifetime mortgages and swaps. Under IAS 39, hedge accounting is only permitted if (amongst other conditions) there is formal designation and documentation of the hedging relationship and of the entity’s hedging strategy and if the hedge is expected to be “highly effective” in achieving offsetting changes in fair value attributable to the hedged risk during the period for which the hedge is designated. The society calculated the carrying value of the lifetime mortgages and the effectiveness of the hedges on the basis that the lifetime mortgages would mature on the same dates as the swaps with which they were paired. This was almost certain to be untrue, particularly where swaps were for terms of 50 years. Plainly, the chance that a borrower aged over 50 when the mortgage loan was made would live (and remain in the same home) for (exactly) the next 50 years was vanishingly small. However, Grant Thornton advised that it was permissible to apply hedge accounting on the basis that, when a mortgage was redeemed before the maturity of the swap with which it was paired, another mortgage loan - at the same fixed rate of interest - could be substituted for it as the hedged item. Whether such substitution would in reality be possible would clearly depend on the level of interest rates prevailing when the mortgage was redeemed.

The society relied on Grant Thornton’s advice in preparing its financial statements for each of the years ending 31 December 2006 to 2011. The society also relied on Grant Thornton’s advice that its use of hedge accounting was legitimate when entering into more lifetime mortgages and swaps during this period.

For each of the years ending 31 December 2006 to 2011, Grant Thornton audited the society’s financial statements and, in each year, signed an unqualified audit opinion certifying that the financial statements gave a true and fair view of the society’s financial position. Each audit report also repeated Grant Thornton’s advice that the society was entitled to apply hedge accounting.

Following the 2008 financial crisis, there was a sustained fall in interest rates which caused the MTM value of the society’s interest rate swaps to become a substantial liability. This liability was offset on the society’s balance sheet by the adjustment made to the reported value of the mortgages using hedge accounting. However, because of the negative MTM value of the swaps, the society was required to provide cash collateral to swap counterparties which, by the third quarter of 2012, amounted to £39.29m.

In March 2013, Grant Thornton informed the society that it was not after all permitted to apply hedge accounting in preparing its financial statements. The effect of that realisation was that the society had to account for the fair value of the swaps in its 2012 accounts without any adjustment to the book value of the mortgages. The society also had to restate its accounts for 2011, with the result that the society’s profit of £6.35m for 2011 became a loss of £11.44m and its net assets were reduced from £38.4m to £9.7m.

As a result of these corrections to its accounts, the society had insufficient regulatory capital. Whereas in 2011 it had reported capital “headroom” of £20.4m, once it was appreciated that hedge accounting could not be used, this figure had to be altered to a capital deficit of £17.9m (ie an overall downwards adjustment of £38.3m).

To extricate itself from this situation, the society terminated all of its interest rate swap contracts early at a cost of £32,682,610. This amount comprised the MTM value of the swaps as at 6 and 7 June 2013, when they were closed out, and transaction costs of £285,460 payable to the swap counterparties because of the early termination of the contracts.

The society also sold its book of UK lifetime mortgages for £68.4m in December 2013. This sale price reflected a small premium (of £3.5m) on the “par” value of the mortgage book (ie the total amount loaned plus the rolled up interest). The society did not sell the Spanish lifetime mortgages because it did not receive an offer at or close to “par”. It was agreed at trial that, on a sale of these mortgages, the society would be likely to receive £3.3m more than their par value - amounting to a net profit of £2.46m after deducting the costs of administering the Spanish mortgage book.

In these proceedings the society has claimed compensation from Grant Thornton for losses suffered as a result of relying on Grant Thornton’s negligent advice. The main loss claimed (and the only head of claim still in issue) is the amount paid to close out the swaps in 2013.

In its defence Grant Thornton admitted that it had been negligent in advising the society when auditing its accounts for each of the years 2006 to 2011 that the society was entitled to apply hedge accounting. There were various reasons why that advice was wrong and should not have been given by any reasonably competent accountant. A major reason was that there was a significant mismatch between the 50-year duration of many of the swaps and the much shorter expected duration of the mortgages. Furthermore, the substitution of mortgages hedged by the swaps on which the society was relying to achieve a match was not permitted under IAS 39.

Grant Thornton defended the claim on the grounds that its negligence did not cause the losses claimed by the society and/or that those losses are not recoverable in law because they are not losses from which Grant Thornton owed the society a duty to protect it.

The trial of the action took place over 17 days in the Commercial Court before Teare J: [2018] EWHC 963 (Comm); [2018] PNLR 27. For reasons given in a careful judgment, the judge awarded damages to the society of only £316,845 (plus interest) - the main item of damages being the transaction costs payable to terminate the swaps early.

The only dispute of primary fact at the trial was whether Grant Thornton was aware in April 2006 that the society intended to hedge lifetime mortgages by entering into long-term swaps. This went to whether Grant Thornton had been negligent in advising the society in April 2006 that hedge accounting could be used, in addition to its admitted negligence in giving such advice when subsequently auditing the society’s accounts for the years ending 31 December 2006 to 2011. The judge decided this issue in favour of the society (see para 71 of the judgment).

The judge also found as a fact that, but for Grant Thornton’s negligent advice in April 2006 and in its subsequent audits, the society would not have entered into any more long-term interest rate swaps after April 2006 and would have closed out those already entered into (para 139). In that event the costs of £32.7m paid to close the swaps in 2013 would not have been incurred. The judge further found that Grant Thornton’s negligent advice was an effective cause in law of that loss (para 146) and that the loss was a reasonably foreseeable consequence of Grant Thornton’s negligence and therefore not too remote to be recoverable (para 201).

Nonetheless, the judge held that the society could not recover damages for this loss from Grant Thornton (apart from the transaction costs paid to terminate the swaps early). He derived from SAAMCO, as explained in Hughes Holland, the test that a defendant is only responsible for losses if they flow from matters for which the defendant has “assumed responsibility” (paras 150-151). The judge considered that there were cogent arguments on each side of this issue, but ultimately concluded that this test was not satisfied essentially because, “looked at broadly, sensibly and in the round,” the losses flowed from market forces for which Grant Thornton did not assume responsibility (para 179).

In case he was wrong in this conclusion, the judge went on to make a number of further findings. In particular:

He rejected a case put forward by Grant Thornton that, even if its advice that hedge accounting could be used had been correct, it is likely that the society would in any event have been forced by the regulator to close out the swaps, incurring the losses that it did in fact incur in 2013 (para 189).

He found that, if Grant Thornton’s advice had been correct, the losses incurred in closing out the swaps would not have been incurred because the swaps would not have been closed out (paras 193-194). The judge did not consider that it was appropriate to expand the inquiry beyond this by asking whether the same losses would have been incurred over the next 30 years or so (the remaining term of many of the swaps).

If, however, it was for the society to prove on the balance of probabilities that the sums paid in 2013 would not have had to be paid over the next 30 or more years, the judge found that the society could not prove this, as “[t]rying to predict what is likely to happen to financial markets, interest rates, the fortunes of the society and the views of the regulator over the next 30 or more years is an impossible task” (para 200).

The judge further found that the society had itself been negligent in two respects which contributed to its loss. The first was its decision to enter into 50-year swaps. This was negligent because the duration of the swaps greatly exceeded the likely duration of the lifetime mortgages and the society could not be confident that it would be able to replace mortgages when they were redeemed with new mortgages on similar terms. The judge noted that the society’s decision to enter into the 50-year swaps was based on assumptions about the level of interest rates in the far distant future, which was an uncertain and speculative exercise, and described the decision as “an unnecessary and imprudent risk to take” (para 239). The second respect in which the society was negligent consisted in the fault of its Finance Director in devising the society’s approach to hedge accounting and considering that hedge accounting was available when it was not (paras 241-242). After evaluating the relative blameworthiness and causative potency of the society’s fault in comparison with that of Grant Thornton, the judge decided that, if (contrary to his conclusion) Grant Thornton was liable to compensate the society for the costs incurred in terminating the swaps, the damages awarded should be reduced by 50% to reflect the society’s contributory negligence (para 255).

The society appealed to the Court of Appeal against the judge’s decision that it could not recover the costs incurred in terminating the swaps (apart from the transaction costs). The appeal was dismissed for reasons given by Hamblen LJ (with whom Males LJ and Dame Elizabeth Gloster agreed): [2019] EWCA Civ 40; [2019] 1 WLR 4610.

The Court of Appeal held that the judge had erred in approaching the issue of liability on the basis of assumption of responsibility and that he should have considered whether Grant Thornton gave “advice” or only “information” to the society (para 59). Applying that distinction as explained by this court in Hughes Holland, the Court of Appeal concluded that, on the undisputed facts and the judge’s findings, this is not an “advice” case where Grant Thornton was responsible for “guiding the whole decision making process” and hence liable for all the foreseeable financial consequences of the decisions to enter into the swaps (paras 63-64); rather, this is an “information” case, such that Grant Thornton is legally responsible only for the foreseeable financial consequences of its information / advice being wrong (paras 54 and 74).

The Court of Appeal further held that the judge had erred in finding that the society had incurred losses on breaking the swaps which would not have been incurred if Grant Thornton’s advice had been correct (as the swaps would not in that case have been terminated). Hamblen LJ observed that the society could not show that it had suffered any loss just by closing out the swaps at their fair value (paras 80-87). To show that it suffered a loss by terminating the swaps when it did, the society would have had to prove that it would have been better off if it had continued to hold the swaps (para 88). However, on the judge’s own findings (see para 66(iii) above), this had not been proved (para 96). The judge had accordingly been correct to dismiss the claim (para 98).

On this appeal the society invites this court to consider the true nature and ambit of the scope of duty principle and, in particular, how this principle applies to the liability of an auditor. Counsel for the society argue that the correct approach is to identify the nature and purpose of the auditor’s duties and to analyse in a qualitative way whether the loss sustained falls within the scope of those duties, rather than to ask what the position would have been if the “information” supplied by the auditor had been correct. They submit that, on a proper analysis, the costs incurred on breaking the swaps were a loss that fell within the scope of Grant Thornton’s duty, being the very kind of loss which Grant Thornton owed a duty of care to protect the society against when it advised that hedge accounting was permissible and that the accounts gave a true and fair view of the society’s financial position. Alternatively, if it is relevant to consider what the counterfactual position would have been if the “information” provided by Grant Thornton had been correct, it is sufficient (as the judge found) that the society would not in those circumstances have suffered the loss that it did, as it would not have terminated the swaps. The society does not also have to prove that it would not have suffered some other equivalent loss at some other time in the future. If there is to be such an enquiry, the burden of proof falls on the defendant.

In order to assess these arguments, it is necessary first of all to identify the relevant legal principles and then to consider how they apply to claims against auditors, both in general and on the facts of this case. The starting point is the leading case of SAAMCO.

The SAAMCO litigation comprised a group of claims brought by mortgage lenders against valuers who had negligently overvalued properties mortgaged as security for loans. In each case the lender had relied on the valuation in deciding whether and how much money to lend on the security of the property and suffered loss when the borrower defaulted on the loan and the property was sold. The background to the litigation was a substantial fall in property prices in the early 1990s. The question which exercised the courts was whether the lender could recover from the valuer its full financial loss suffered as a result of making the mortgage loan in reliance on the valuation, or whether the damages awarded should be limited in some way to reflect the fact that the losses suffered were partly attributable to the fall in the property market.

In Banque Bruxelles Lambert SA v Eagle Star Insurance Co Ltd [1995] 2 All ER 769, Phillips J held that the lenders could not recover from the negligent valuers “that part of their loss which is attributable to a fall in the collapse of property market” because “[i]t does not seem to me that such loss can fairly and reasonably be considered as resulting naturally from [the valuers’] failure to report as they should have done” (see p 806). He continued:

The Court of Appeal reversed this decision: Banque Bruxelles Lambert SA v Eagle Star Insurance Co Ltd [1995] QB 375. In a judgment delivered by Sir Thomas Bingham MR the court held that the fall in the property market was not unforeseeable and could not be regarded as a new intervening cause of loss, and that in these circumstances the valuer’s negligence was the effective cause of the lender’s entire loss. The court accepted that the valuer was not responsible for the lender’s business investment and, had it valued the property competently, would have had no liability for any loss suffered as a result of a market fall. However, the court held that, once it was shown that the valuer’s negligence had caused the lender to make a loan which it would not otherwise have made and which it was locked into, the valuer was liable for the lender’s full loss, including all the loss attributable to a fall in the property market (see pp 420-421).

In response to the point that it was unfair to make the valuer who had advised on only one head of risk in effect the underwriter of the loan, the Court of Appeal gave a “swings and roundabouts” argument (at p 421G):

Following the Court of Appeal’s decision, the Banque Bruxelles case itself was settled. But appeals in three linked cases decided at the same time by the Court of Appeal proceeded to the House of Lords, whose decision is reported under the name of South Australia Asset Management Corpn v York Montague Ltd [1997] AC 191 (“SAAMCO”). The House of Lords adopted a different approach from either of the courts below, holding that the damages recoverable by the lenders were limited to the amount by which the security was overvalued. The reasons for this conclusion were given by Lord Hoffmann, with whom the other members of the appellate committee agreed.

Lord Hoffmann’s reasoning started from the very general principle that a person who owes a duty to another (whether in contract or tort or under a statute) is not normally liable for all the consequences of a breach of duty but only for losses of the kind in respect of which the duty was owed: see [1997] AC 191, 211H-212F. As authority for this principle, Lord Hoffmann quoted Lord Bridge of Harwich in Caparo Industries Plc v Dickman [1990] 2 AC 605 at 627:

The question raised in Caparo was whether, in auditing a company’s accounts, the auditors owed a duty of care to investors who relied on the accuracy of the information stated in the accounts in deciding whether to buy shares in the company. The House of Lords held that the auditors did not owe such a duty. The essential reason was that the purpose of a statutory audit is to provide the company and its shareholders with accurate information about the company’s finances on which to base management and governance decisions. It is not a purpose of the audit to provide information on which potential investors may rely in deciding whether to buy shares in the company. It made no difference in this regard whether the potential investor was an existing shareholder. Hence where a shareholder suffered loss as a result of relying on the auditors’ opinion in deciding to buy more shares, this was not a loss of the kind which the auditors owed a duty of care to protect the shareholder against.

The scope of the duty of a valuer can likewise, as Lord Hoffmann indicated (at p 212D-G), be identified by analysing the purpose of the service which the valuer undertakes to provide. From the known purpose of a valuation, Lord Hoffmann deduced that a valuer who enters into a contract to provide a valuation owes a duty of care. As for the scope of the duty, “in the sense of the consequences for which the valuer is responsible” (p 212E), he observed that there is no reason in principle why the law should not penalise wrongful conduct by shifting onto the wrongdoer the whole risk of consequences which would not have happened but for the wrongful act. Normally, however, the law does not “make the wrongdoer liable for all the consequences of his wrongful conduct” but “limits liability to those consequences which are attributable to that which made the act wrongful” (see p 213C). Lord Hoffmann illustrated this point by reference to a case discussed by Hart and Honoré in their classic work on Causation in the Law, 2nd ed (1985), pp 118-120: The Empire Jamaica [1955] P 52; affirmed [1955] P 259 (CA); [1957] AC 386 (HL). In that case a collision at sea was caused by the negligence of a ship’s mate who lacked a certificate of competence required by the Merchant Shipping Acts. In proceedings to limit the liability of the vessel’s owners for damage caused by the collision, it was held that the owners had succeeded in proving that the damage occurred without their “actual fault or privity” despite their failure to employ a certificated mate. This was because, on the evidence, the owners had good reason to consider the mate to be competent and there was no causal connection between the lack of a certificate (in respect of which the owners were at fault) and the damage which occurred.

Hart and Honoré pointed out that the requirement to show a causal connection between an aspect of the defendant’s conduct which made it wrongful and the injury suffered by the claimant can be justified in various ways. One way, in a case such as The Empire Jamaica, is to say that the absence of a certificate was causally irrelevant. This in turn can be defended by making a counterfactual argument that, had the mate had a certificate but all other conditions remained the same, the same damage would still have occurred.

Lord Hoffmann applied this analysis to the case of liability in negligence for providing inaccurate information. He gave the now well-known example of a doctor who negligently advises a mountaineer about to undertake a difficult climb that his knee is fit for the task. The mountaineer goes on the climb, which he would not have undertaken if the doctor had told him the true state of his knee and suffers an injury which is “an entirely foreseeable consequence of mountaineering but has nothing to do with his knee”. Lord Hoffmann reasoned that the doctor is not liable for the injury as “[t]he injury has not been caused by the doctor’s bad advice”. He justified this by saying that the injury “would have occurred even if the advice had been correct”: see [1997] AC 191, 213F.

Lord Hoffmann then generalised this reasoning (at p 214C-D) to formulate the following principle:

The novel step in the decision was to apply this principle to divide what was on its face a single financial loss. The Court of Appeal had regarded it as “commercially unrealistic to seek to separate the risk of negligent overvaluation and the risk of a fall in the market” and to ascribe different elements of the loss to different causes, stating: “It is one transaction and one loss” (see [1995] QB 375 at 407). However, the House of Lords performed just such a separation. As Lord Hobhouse later observed in Platform Home Loans Ltd v Oyston Shipways Ltd [2000] 2 AC 190, 209G, the development of the scope of duty reasoning in SAAMCO was that Lord Hoffmann, “instead of applying it to kinds or categories of damage, applied it to the quantification of damage” (emphasis in the original). To quantify the part of the lender’s loss which fell within the scope of the valuer’s duty, Lord Hoffmann applied the test of asking what element of the loss suffered by the lender as a result of lending on the security of the negligently overvalued property was attributable to the inaccuracy of the valuation (see p 216E). That amount could be ascertained by asking what loss would have occurred if the valuation given by the valuer had been correct.

The decision in SAAMCO, although perceived at the time by practitioners and academic commentators alike as being of great importance, gave rise to difficulty and a good deal of controversy in understanding its rationale and implications. This may have been partly due to the fact that aspects of Lord Hoffmann’s reasoning were compressed and, in particular, did not spell out the reasons of policy underlying his view (stated in the passage quoted at para 83 above) that a duty of care which imposes on a professional adviser “responsibility for losses which would have occurred even if the information which he gave had been correct” is not “fair and reasonable as between the parties”.

To identify those reasons, it is necessary to return to the purpose for which a lender commissions a valuation and the role which the valuation is reasonably expected to play in the lender’s business decision. The purpose of the valuation is to provide the lender with an opinion on which it is entitled to rely of the current market value of the property offered as security for the loan. Clearly, the value of the security is an important consideration for a mortgage lender. It is, however, by no means the only factor relevant to the decision whether to make the loan. The lender will also need to assess the credit risk in lending to the particular borrower - a matter for which the valuer has no responsibility. In addition, the valuer is normally asked to assess only what the property is currently worth and not to forecast what it will be worth at a future date when the lender may need to enforce the security. As Lord Hoffmann said in SAAMCO at p 210F: “a valuer provides an estimate of the value of the property at the date of the valuation. He does not undertake the role of a prophet”.

It is obvious that the value of the property mortgaged as security for the loan may subsequently go up or down. The risk that the value of the property will go down is a commercial risk which the lender takes. That does not mean that the lender’s willingness to take this risk is unqualified. The lender may only be willing to take this risk on the understanding that the property is currently worth what the valuer advises it is worth: that necessarily follows where the lender proves that, had the property not been overvalued, it would not have made the loan. But what can be inferred from the fact that the lender did in fact make the loan is that the lender was willing to bear the risk (without relying in this regard on the valuer) that the property would in future be sold for less than the valuation figure in so far as this would have been so even if the valuation had been accurate. To that extent, any loss suffered by the lender can fairly be said to be a consequence of risks inherent in the lending transaction, including the risk of a fall in property prices, and not of the only risk for which the valuer can fairly be held responsible, namely, the risk that the valuation was overstated.

This is the underlying policy rationale for not shifting onto the valuer all the risks taken by the lender in making the loan and instead leaving the lender to bear the risk of loss which would have occurred even if the valuation had been correct. The aim is to allocate responsibility for any loss incurred by the lender in a way which fairly reflects the assumption of risk implicit in the service which the valuer agreed to provide.

This policy rationale was articulated by Lord Nicholls of Birkenhead in Nykredit Mortgage Bank Plc v Edward Erdman Group Ltd [1997] 1 WLR 1627 at 1631, when the House of Lords dealt with the question of what interest was payable on the damages awarded in SAAMCO. Lord Nicholls explained that:

In SAAMCO Lord Hoffmann drew a distinction between “a duty to provide information for the purpose of enabling someone else to decide upon a course of action and a duty to advise someone as to what course of action he should take”: see [1997] AC 191, 214E (emphasis in the original). The limitation on the consequences for which the defendant is potentially liable was said to apply only in the former category of case, where the duty is to provide information. By contrast:

In Hughes-Holland, at para 39, Lord Sumption acknowledged the “descriptive inadequacy” of the labels used - in that information given by a professional person to a client “is usually a specific form of advice”, and “most advice will involve conveying information. Neither label really corresponds to the contents of the bottle”. Lord Sumption made it clear that the relevant distinction is based, not on the nature of the defendant’s work product and whether that is more naturally described as “information” or “advice”, but solely on the extent of the responsibility undertaken by the defendant. What characterises cases falling within Lord Hoffmann’s “advice” category is that:

Two points arise from this. First, rather than continuing to use labels which are misleading, it seems to me that it would be desirable to dispense with the descriptions “information” and “advice” as terms of art and to focus instead on the need to identify with precision in any given case the matters on which the professional person has undertaken responsibility to advise and, in the light of those matters, the risks associated with the transaction which the adviser may fairly be taken to owe a duty of care to protect the client against. What determines whether the adviser has a duty to protect the client against the full range of risks associated with a potential transaction, or only against some of those risks, is whether or not the adviser’s contribution to the decision-making process is limited. If it is not, either because the adviser is responsible for recommending a course of action (and not just for providing material relevant to the client’s decision as to what course of action to take) or because the matters which the adviser is expected to take into account are open-ended and not limited to a particular subject matter, then the adviser’s responsibility extends to all the foreseeable risks of entering into the transaction. Conversely, if the adviser’s contribution is limited to advising on particular matters or a particular sub-set of considerations relevant to the client’s decision, the risks which the adviser has a duty to protect the client against will be correspondingly limited.

It follows from the nature of this distinction that cases in which a professional adviser is liable for all the foreseeable consequences of a commercial transaction entered into as a result of negligent advice are likely to be rare. The position may be different where, for example, the claimant is a private individual relying on a financial adviser to recommend an investment. But in a commercial context it is unusual for a professional adviser to be asked to advise on the overall merits of a transaction or left to decide on the matters to consider in formulating their advice. It will usually be clear that the adviser’s responsibility is limited to a particular area of expertise and that there will be other considerations relevant to the client’s decision which are not for the adviser to assess. As Lord Millett observed in Aneco Reinsurance Underwriting Ltd (in liquidation) v Johnson & Higgins Ltd [2001] UKHL 51; [2002] 1 Lloyd’s Rep 157, para 62, in identifying the area of responsibility of a professional adviser, his or her profession will usually supply the answer.

A second point clarified in Hughes-Holland is that the distinction between cases where the adviser’s responsibility extends to all the foreseeable risks associated with the client’s decision and cases where it is limited to a particular risk or area of risk does not depend on how important the adviser’s contribution is to the decision taken. Where the material contributed by the defendant is limited in scope, the defendant’s responsibility is correspondingly limited, “even if the material which the defendant supplied is known to be critical to the decision to enter into the transaction” (para 41). Lord Sumption pointed out that, if it were otherwise, the defendant would be potentially liable for the entire foreseeable loss flowing from the transaction in every case where the claimant would not have entered into the transaction but for the defendant’s advice. That, however, was precisely the proposition rejected in SAAMCO (see para 42). For this reason, the Supreme Court disapproved two decisions in which SAAMCO had been (wrongly) distinguished: Bristol and West Building Society v Steggles Palmer reported with Bristol and West Building Society v Fancy & Jackson [1997] 4 All ER 582; and Portman Building Society v Bevan Ashford [2000] PNLR 344. In those cases the defendant solicitors were held to be liable for the whole loss flowing from mortgage loans because they negligently failed to provide the lender with information which was fundamental to its decision whether to enter into the transaction. Lord Sumption observed that this is true ex hypothesi in every case where the test of factual causation is satisfied. To say that a fact misrepresented or not reported was critical or fundamental to the decision, or would have shown that the transaction was not viable, are all simply different ways of saying that, if the claimant had not received the advice it did, it would not have entered into the transaction: see paras 51-52. Whether in such circumstances the defendant is liable for all foreseeable loss flowing from the transaction depends not on the gravity or causative potency of the defendant’s error or omission but on the scope of the matters for which the defendant undertook responsibility.

In SAAMCO Lord Hoffmann described the consequences for which an adviser whose duty of care is limited in scope is responsible, if negligent, as “all the foreseeable consequences of the information being wrong”: see [1997] AC 191, 214F. If read out of context, this description could be, and has sometimes been, mistaken to mean that the adviser is liable for the consequences of providing wrong information. It is clear, however, that this is not what Lord Hoffmann meant, as again this would simply be another way of stating the requirement to show factual causation. In any case where the claimant has entered into a transaction as a result of the provision of incorrect information or advice, all the consequences of entering into the transaction are consequences that would not have occurred if correct advice had been given. The whole point of the decision in SAAMCO is that the adviser is not liable for all such consequences, but only for those which arise from matters within the adviser’s area of responsibility.

The relevant causal relationship for this purpose is not between the provision of information or advice and the claimant’s loss but between what made the information or advice wrong and the loss. What makes information or advice wrong is the existence of facts or matters which the adviser has misrepresented or failed to report. It is the foreseeable consequences of those matters to which the adviser’s responsibility is limited.

This can be illustrated by considering Lord Hoffmann’s example of the mountaineer’s knee. If the question asked is whether the injury suffered by the mountaineer was a consequence of the doctor’s negligent advice, the answer is that it was - in that, if the doctor had given correct advice that his knee was unfit, the mountaineer would not have gone on the expedition and hence would not have suffered the injury that he did. If the focus is on the consequences of the doctor’s advice, the only way of limiting the extent of the doctor’s responsibility for the injury is by invoking concepts of remoteness or legal causation. It might be said, for example, that although the doctor’s advice that the knee was sound was a factual (or “but for”) cause of the injury, it was not an “effective” cause. For the purpose of determining whether the injury was within the scope of the doctor’s duty, the relevant question is different. The question here is not whether there is a sufficient causal connection between the injury and the advice given by the doctor, but whether there is a sufficient causal connection between the injury and the state of affairs which made the advice incorrect, that is to say the condition of the knee. On the assumed facts, the injury sustained by the mountaineer had nothing to do with the state of his knee (see para 82 above). It was therefore not within the scope of the doctor’s duty of care.

I think that it is this distinction which Lord Hoffmann had in mind when he said in Nykredit (at p 1638F) that the extent of the valuer’s liability “has nothing to do with questions of causation”. This may seem a surprising remark given that the ratio of SAAMCO is that a valuer who negligently overvalues property mortgaged as security for a loan is liable only for loss suffered by the lender insofar as the loss is shown to have been caused by the inaccuracy of the valuation. Identifying whether or to what extent the claimant’s loss is a consequence of the information provided by the valuer being wrong (in the sense that the security is worth less than the valuer advised that it was worth) is manifestly a question of causation. What Lord Hoffmann meant by his remark can, however, be seen from the passage (at p 1638G-H) which immediately followed it:

In other words, the decision of the House of Lords in SAAMCO was not concerned with the questions which arise in working out whether the defendant’s negligent conduct foreseeably caused loss to the claimant. In SAAMCO that element of the claims was admitted. What the House of Lords held was that the valuer was not responsible for the whole loss foreseeably caused by the negligent valuation advice but only for such part of the loss as was a consequence of the lender having less security than the valuer had represented. Asking what part of the lender’s loss is a consequence of the lender having too little security is obviously a question of causation. But it is a different question, which it is important to keep distinct, from the questions of whether the defendant’s negligent advice was a factual and legal cause of the claimant’s loss. Lord Sumption made this point in Hughes-Holland, when he said (at para 38):

A further aspect of Lord Hoffmann’s reasoning in SAAMCO which has given rise to difficulty is his use of a counterfactual test to identify the loss for which a negligent provider of “information” (in his terminology) is responsible, by posing the question: would the loss suffered by the claimant have occurred even if the information provided by the defendant had been correct? I have sought to explain the policy rationale underlying this test. The test has, however, sometimes been misunderstood (including, as I discuss below, in the present case). It is also sometimes difficult to apply.

One source of difficulty is the intrinsic vagueness of counterfactual propositions: see David Lewis, Counterfactuals (1973) Chapter 1; Michael S Moore, Causation and Responsibility: An Essay in Law, Morals and Metaphysics (2009) Chapter 16. In order to yield a determinate answer to a counterfactual question, assumptions need to be made about how precisely the counterfactual world is supposed to differ from, or remain similar to, the actual world. This can be seen in SAAMCO itself. Ascertaining what loss would have occurred if a negligent overvaluation had been accurate can be approached in various ways. In SAAMCO several different approaches were suggested by the valuers (discussed at pp 219F-221E). The approach preferred by the House of Lords was to limit the recoverable loss to the amount by which the property was overvalued. This amount operated, in effect, as a cap on the lender’s damages and is sometimes referred to as the “SAAMCO cap”.

Criticisms have been made of this approach. It assumes that the amount of additional security which the lender believed that it had over and above the true value of the security at the time of the valuation would have remained intact when the market fell and the property came to be sold. Commentators have argued that it would be fairer and more realistic to assume that the value of this additional security would have depreciated at the same rate as the value of the actual property or, alternatively, at the same rate as the value of a comparable property which at the time of the valuation was worth what the valuer represented the value of the actual property to be: see Edward Davidson, “BBL and damages: some problems in applying the ratio decidendi” (1997) 13 PN 89, 92-93; DW McLauchlan “A Damages Dilemma” (1997) 12 JCL 114, 128; John Murdoch, “Negligent Valuers, Falling Markets and Risk Allocation” [2000] 8 Tort Law Rev 183, 202-203; Nick Hegan, “SAAMCO, The Scope of the Duty and Liability for Consequences” (2007) 38 Victoria University of Wellington Law Review 465, 469-470. Another, more complicated approach, described as the “constructive actualised discount method”, is proposed by Alexander Loke, “The valuer’s liability for negligent valuation - toward a more principled allocation of the risk of market decline” (1999) 19 Legal Studies 47. What all these methods have in common is that, unlike the method adopted in SAAMCO, they take account of the fact that the effect of the overvaluation on the lender’s loss is not fixed in aspic when the valuer’s advice is given but depends on subsequent events.

I can illustrate this point with an example of a property negligently valued at £10m when its true value was £8m. The method of quantification adopted in SAAMCO would set a “cap” on the damages recoverable by the lender of £2m (the amount of the overvaluation). This same cap would apply however much or little the property subsequently fell in value. By comparison, if the market collapsed such that the value of the property fell by 50% (ie from £8m to £4m), the first of the alternative methods mentioned above would assume that, if the valuation figure (of £10m) had been correct, the value of the property would similarly have fallen by 50% (to £5m). Using this method, the maximum recoverable loss (ie the amount attributable to the valuation being wrong) would therefore be £1m rather than £2m.

There is force in the criticisms made of the particular method used in SAAMCO. In Hughes Holland, at para 46, Lord Sumption acknowledged that the “SAAMCO cap” is “mathematically imprecise” but defended it by citing Lord Hobhouse’s observation in Platform Home Loans [2000] 2 AC 190 at 207, that the formula is “essentially a legal rule which is applied in a robust way without the need for fine tuning or a detailed investigation of causation.”

The present appeal is not the occasion to revisit the question whether the particular method employed in SAAMCO is the best one to use to quantify the loss recoverable by the lender in a case involving the negligent overvaluation of security. The important point for present purposes is more general. Where a counterfactual test is used, this should not be seen as a mechanical exercise. Care is needed to seek to ensure that the assumptions adopted are suitable to reflect the allocation of risk between the parties to which the test is designed to give effect. There are also cases in which a counterfactual test cannot readily be applied. Some possible examples of such situations are given in an article by Hugh Evans, “Solicitors and the scope of duty in the Supreme Court” (2017) 33 PN 193. One is a case where a solicitor advising a lender negligently fails to discover and report that the borrower is a former bankrupt (see Omega Trust Co Ltd v Wright Son & Pepper (No 2) [1998] PNLR 337). Had the lender known this information, it would not have made the loan. If the borrower later fails to repay the loan, then, in deciding whether the loss sustained by the lender is within the scope of the solicitor’s duty, it does not make much sense to ask whether, if the (implied) advice that the borrower had not previously been bankrupt had been true, the borrower would have repaid the loan so that the loss would not have occurred. Nevertheless, the previous bankruptcy clearly indicated that lending to that borrower involved a significant credit risk, which it was part of the solicitor’s duty to take care to protect the lender against. It is not difficult to conclude, without the need to use counterfactual reasoning, that there was a sufficient causal connection between the subject matter of the duty and the borrower’s default to regard the loss suffered by the lender as the materialisation of that credit risk and therefore within the scope of the solicitor’s duty.

Accordingly, it should not be assumed that it is necessary or helpful to apply a counterfactual test of the kind formulated by Lord Hoffmann in every case. Such a test is simply one means of assessing whether there is a sufficient causal connection between the subject matter of the defendant’s advice and the loss suffered by the claimant to justify the conclusion that the loss arose from a risk which was within the scope of the defendant’s duty. Causal connections are not always amenable to analysis in counterfactual terms. Without seeking to be definitive or exhaustive, there seem to me to be two situations in which a counterfactual test is likely to be most useful. One is where, as in the valuers’ negligence cases, it is necessary to divide into separate elements what on its face is a single loss. In such cases, as discussed above, a counterfactual test can provide a method for quantifying what part of the loss is the materialisation of a risk which the defendant owed a duty of care to protect the claimant against and what part of the loss is attributable to other risks which the claimant would have taken upon itself even if the defendant’s advice had been correct. The other situation in which a counterfactual analysis is conspicuously useful is to help explain why - where this is the case - the subject matter of the defendant’s negligent advice (or other wrongdoing) was causally irrelevant to the injury which occurred - as in The Empire Jamaica, or Lord Hoffmann’s hypothetical example of the mountaineer’s knee or, I would add, the case of Khan v Meadows.

Apart from the particular method used to quantify the “SAAMCO cap”, the principles for which SAAMCO is authority are clearly not limited to cases involving negligent valuations and are of much broader application. Since SAAMCO was decided, it has become commonplace for professional advisers resisting claims for damages to argue that, even if they were negligent and the claimant suffered loss as a result of their negligent act or omission, the loss is not recoverable as it was not within the scope of the adviser’s duty of care. It is unnecessary for present purposes to consider the extent to which the same or similar principles apply outside the context of professional liability for giving negligent advice. I will, however, mention the two cases in which the House of Lords or this court has applied the principles discussed in SAAMCO to professional advisers other than valuers.

Aneco Reinsurance Underwriting Ltd (in liquidation) v Johnson & Higgins Ltd concerned the liability of a firm of insurance brokers who offered to arrange reinsurance cover for a risk which they were seeking to persuade the claimant underwriter to accept. The brokers reported that they had placed the reinsurance. However, they had failed to present the risk fairly to the reinsurers, with the result that the reinsurance cover was later avoided. On the facts it was found that, if the brokers had presented the risk fairly, it would have been impossible to obtain cover in the market; and also that, without the reinsurance cover, the claimant would not have agreed to underwrite the inwards business. The question was whether the claimant could recover from the brokers its entire loss (of $35m) on the business underwritten in reliance on the brokers’ negligent misrepresentation that (valid) reinsurance cover had been obtained; or only the amount (of $11m) that would have been recoverable under the reinsurance if it had been valid. By a majority of 4-1, the House of Lords held that the claimant could recover its entire loss.

The decision turned on the correct analysis of the extent of the duty which, on the facts, the brokers had undertaken towards the underwriter. The majority approved the view of Evans LJ in the Court of Appeal that the brokers had undertaken a responsibility, not just to arrange reinsurance or report that it was unavailable, but, in circumstances where no reinsurance cover could be obtained if the risk was fairly presented, to advise the underwriter about the unfavourable market assessment of the risk which they were seeking to persuade him to underwrite. On this basis the majority concluded that the scope of the brokers’ duty extended to advising the underwriter about the wisdom of entering into the whole transaction, with the result that the brokers were liable for the full loss which flowed from doing so. Lord Millett, who dissented, made cogent criticisms of this analysis involving, as it did, an inference that a broker had undertaken responsibility for an underwriting decision (see para 66(4)). However, he agreed that, if that were the correct analysis of the facts, the full loss suffered by the underwriter would be attributable to the brokers’ breach of duty (para 92).

The more recent decision of this court in Hughes-Holland v BPE Solicitors was a case of solicitors’ negligence. The defendant solicitors were instructed to document a loan made by their client, Mr Gabriel, to a developer, Mr Little. Mr Gabriel understood that the money would be used to finance a property development. In fact, Mr Little used the loan to repay another loan secured on the property and to meet other liabilities; no development of any significance was carried out, the property was worthless and Mr Gabriel lost all his advance. In Mr Gabriel’s subsequent claim against his solicitors, the trial judge found that they had been negligent in using a form of loan agreement which inadvertently confirmed Mr Gabriel’s assumption that the loan money would be used to finance the development. Although the judge awarded damages to Mr Gabriel, that decision was reversed by the Court of Appeal whose decision was affirmed by the Supreme Court. (By the time of the Supreme Court hearing Mr Gabriel was bankrupt and had been replaced as claimant by his trustee in bankruptcy, Mr Hughes-Holland.)

Lord Sumption, with whose judgment the other members of the court agreed, explained (at paras 54-55) that the claim failed because the solicitors had only been instructed to prepare the loan documentation and not to advise on the decision to make the loan. Their liability arose only from having negligently confirmed (by using the wrong standard form) Mr Gabriel’s assumption about how the loan was to be used. The Supreme Court thought it clear from the evidence that, even if the money had been used to finance the development, the development would not have been completed and the property would have remained worthless. Mr Gabriel would therefore still have lost his money. It followed that the loss was not caused by a risk (that the loan might be used for other than its intended purpose) which the solicitors could be said to have owed a duty of care to protect Mr Gabriel against.

In these two cases, the critical part of the analysis was to identify the scope of the defendant’s duty. Once that had been done, it was reasonably straightforward to determine whether the loss suffered by claimant was within the scope of that duty. In cases of auditor’s negligence, identifying the scope of the auditor’s duty does not usually give rise to difficulty. However, determining whether loss falls within the scope of the auditor’s duty of care is not always straightforward.

Like other professional advisers, an accountancy firm acting as an auditor owes a duty, as an implied term of the contract by which the firm is retained and also in tort, to carry out the service which it has agreed to supply with reasonable care and skill. As mentioned earlier and discussed in detail in Caparo Industries Plc v Dickman, the purpose of a statutory audit is to seek to ensure that the audited entity and its shareholders are provided with accurate information about the company’s financial position on which to base management and governance decisions. In Caparo Lord Oliver of Aylmerton (at p 630F-G) explained the purpose of the auditor’s duty as follows:

Provided that the audited entity and its members are supplied with accurate financial information, what they do with this information is a matter for them and falls outside the scope of the statutory purpose: Berg Sons & Co Ltd v Adams [1992] BCC 661 at 677 (Hobhouse J). The responsibility of the auditor does not extend to advising on the conduct of the audited entity’s business or the commercial merits of any transaction. As Lindley LJ stated in In re London and General Bank (No 2) [1895] 2 Ch 673 at 682:

It thus follows from the purpose of a statutory audit that the auditor’s duty of care is limited to protecting the audited entity and its members against the risk that its audited accounts are inaccurate (together with the risk of certain types of wrongdoing such as, in Lord Oliver’s example, paying a dividend out of capital). It is no part of the auditor’s duty to advise the audited entity what business decisions it should make nor to identify what considerations apart from the accuracy of its audited accounts are relevant for the entity to take into account in making such decisions. Those are exclusively matters for the entity’s directors and members (or possibly other advisers) to assess. Applying the principles established in SAAMCO and Hughes-Holland discussed at paras 90-94 above, cases of auditor’s negligence are therefore a classic instance of the situation in which the professional adviser does not owe a duty of care to protect the client against all the foreseeable risks and consequences of transactions entered into by the client in reliance on the advice given, but only against those which are related to the subject matter of the advice. This means that the auditor is liable only for losses caused by matters which the auditor negligently misstated or failed to detect or report, and not for losses unconnected with those matters which the entity would have suffered even if the auditor’s advice had been correct.

Although decided before SAAMCO and therefore without reference to it, Galoo Ltd v Bright Grahame Murray [1994] 1 WLR 1360 provides a good example of a claim to recover damages for losses which, applying this test, could not be said to be within the scope of the auditor’s duty of care. On the facts alleged, the defendant auditors had negligently failed to discover and report that the value of stock and work in progress shown in the accounts of the claimant companies was substantially overstated. If correctly stated, the accounts would have revealed the companies to be insolvent. This in turn would have led them to cease trading. As it was, as a result of the auditor’s negligent advice that the accounts gave a true and fair view of their financial position, the companies continued to trade and, in doing so, incurred further losses. The Court of Appeal held that the claim to recover those losses had rightly been struck out as, on the facts alleged, the auditors had not caused the claimants’ trading losses but had only provided the opportunity for those losses to be incurred.

To justify this conclusion, Glidewell LJ (with whom Evans and Waite LJJ agreed) relied on “the application of the court’s common sense”: see [1994] 1 WLR 1360, 1375. A mere appeal to common sense, however, is not a satisfactory explanation for the decision, not least since (as Glidewell LJ had himself observed at p 1372) “inevitably, not all judges regard common sense as driving them to the same conclusion” - or, as Langley J put it more bluntly in Equitable Life Assurance Society v Ernst & Young [2003] Lloyd’s Rep PN 88 at para 85: “One person’s common sense may be another’s nonsense”.

As Langley J suggested and as Lord Hoffmann pointed out in a lecture to the Chancery Bar Association, “Common Sense and Causing Loss” (15 June 1999), the decision in Galoo is better explained in terms of the scope of duty principle. In order to show that the losses suffered by the companies in continuing to trade were losses which the auditors owed a duty to protect the companies against, it would have been necessary to plead and prove the existence of a causal link between the losses and those matters which the auditors negligently failed to detect and which made the accounts inaccurate - namely, the fact that the companies’ stock and work in progress was worth much less than the values shown in the accounts. However, the claimants did not allege that there was any such causal link. The claim was advanced simply on the basis that the negligence of the auditors caused loss because, and solely because, it allowed the companies to continue to trade. In those circumstances, even if the facts alleged were proved, the claim could not succeed as those facts were not sufficient to show that the trading losses were within the scope of the auditors’ duty of care. In counterfactual terms, no attempt was made to disprove the possibility that the losses would have occurred even if the value of stock and work in progress shown in the accounts had been accurate. The claim was therefore rightly struck out.

By contrast, in Barings Plc v Coopers & Lybrand (No 7) [2003] EWHC 1319 (Ch); [2003] Lloyd’s Rep IR 566 there was a connection between the matters which the auditors negligently failed to detect and report (unauthorised trading conducted by the general manager of a company in the Barings group) and subsequent losses caused by the continuation of such unauthorised trading. The losses were thus attributable to a risk which was within the scope of the auditors’ duty of care.

Similar reasoning seems to me to justify the recent decision of the Court of Appeal in Assetco Plc v Grant Thornton UK LLP [2020] EWCA Civ 1151; [2021] Bus LR 142; [2021] PNLR 1. In that case the managers of a holding company dishonestly prepared accounts which presented an entirely false picture of the group’s finances. The accounts showed the company and the group to be successful and profitable when in truth the group was insolvent. The auditors negligently failed to detect the falsity of the accounts and gave an unqualified audit opinion for two consecutive years. During those two years, the group carried on with two contracts which, although presented in the accounts as profitable, were in fact heavily loss-making. By doing so, the group incurred further substantial losses. When the true state of affairs became apparent, new managers were appointed, the loss-making contracts were terminated and the group was rescued by an investment of further capital. The trial judge found that, if the auditors had uncovered the true position two years earlier as they should have done, the same measures would have been taken then and the losses sustained in the meantime would have been avoided.

The judge held that the auditors were liable for most of those losses (subject to a 25% reduction for the company’s contributory negligence). The principal ground of appeal was that the losses were outside the scope of the auditors’ duty of care. Save in relation to one particular transaction, this argument was rejected by the Court of Appeal. David Richards LJ (with whom Phillips LJ and Sir Stephen Richards agreed) held that the losses fell within the scope of the auditors’ duty because their negligence deprived the company of information that would have caused it to cease its loss-making activities and to take the steps necessary to regain its solvency.

That reasoning on its own is unsatisfactory, as it amounts to saying only that, but for the auditors’ negligence, the losses would not have occurred. This is merely the ordinary test of factual causation. To conclude that the losses were within the scope of the auditors’ duty, it was necessary to go further and find that the losses arose from a matter which made the accounts and auditors’ opinion on the accounts incorrect.

Although not clearly spelt out in the judgment of the Court of Appeal, however, it appears that this requirement was indeed met. Where an auditor negligently fails to detect and report a cause or potential cause of loss to the audited entity, with the result that losses from that cause occur or continue to occur, there is ex hypothesi a causal link between the subject matter of the auditor’s negligence and those losses. Thus, on the facts of Assetco, it seems clear that the two contracts which the auditors negligently failed to identify as heavily loss-making were inherently likely to continue to generate further losses, as in fact happened, unless or until they were terminated. The losses on the contracts were in those circumstances losses which the auditors owed a duty of care to protect the claimant against.

We were also referred to Deloitte & Touche v Livent Inc 2017 SCC 63; (2017) 416 DLR (4th) 32, in which the Supreme Court of Canada was split 4-3 on a question whether the “SAAMCO principle” precluded an auditor’s liability for trading losses. The difference of opinion seems largely to have turned on how the facts of that case should be analysed. However, counsel for the society cited the following statement of the “SAAMCO principle” by the majority of the court (at para 90):

On this appeal counsel for the society have argued that the “SAAMCO cap” is not applicable to cases of auditors’ negligence. If by the “SAAMCO cap” they had merely meant the particular method adopted in SAAMCO to quantify the loss caused by the overvaluation of the security, I would agree that (as discussed at paras 102-104 above) the “SAAMCO cap” is a blunt instrument which is not suitable to be applied more widely. It is clear, however, that the society’s argument was directed not merely to the particular method adopted in SAAMCO to give effect to the proposition that the adviser is not liable for losses which would have occurred even if its advice had been correct, but to that proposition itself. Counsel for the society submitted that this counterfactual test should not be regarded as a “universal panacea” and that “[l]ike any tool it has its uses but also its limitations.” With that, I agree for the reasons indicated at paras 105-106 above. They also submitted, however, more broadly, that the test is not suitable for deployment in auditor’s negligence cases. The argument advanced in support of this contention is, in my view, ill-founded for reasons which show the importance of a correct understanding of the counterfactual test and its rationale.

Counsel for the society observed that the usual complaint in an audit case is that, as a result of the auditor’s negligence, the client assumed that the financial position set out in its audited accounts was correct and acted accordingly - precisely as it would have done if the stated information had in fact been correct. They submitted that limiting recoverable damages to reflect the position that the client would have been in if the information had been correct will, in consequence, “result in a nil claim: the counterfactual will just replicate the position that the client is actually in”.

As an example, counsel for the society posited a hypothetical case in which a company’s audited accounts show a profit of £50m when, on a true and fair view, its profit was nil. In reliance on the auditor’s report, the company pays a dividend of £25m. It is established law that in such a case, where a dividend is paid out of capital and not out of distributable profits, the company (at least where it is insolvent) is entitled to recover the amount of the dividend as damages from its auditor: see eg Leeds Estate, Building and Investment Co v Shepherd (1887) 36 Ch D 787, 809; In re London and General Bank (No 2) [1895] 2 Ch 673, 686-688. However, counsel for the society submitted that limiting the auditor’s liability to loss that would have occurred if the auditor’s advice had been correct would defeat that claim, since, if the advice had been correct, the company would still have paid out the dividend and suffered the same loss. They argued that the fact that such a conclusion is plainly wrong shows that the counterfactual question articulated in SAAMCO is not applicable to auditor’s negligent cases.

I agree with the submission of Mr Salzedo QC for Grant Thornton that this argument betrays a misunderstanding of the principle. As explained earlier, the point of the counterfactual test is not to consider whether, if the advice given by the defendant had been correct advice to give, the claimant would have acted differently (which self-evidently it would not). It is to consider whether, if the advice had been correct in the sense that the facts had been as the defendant represented them to be, the claimant’s action would still have resulted in loss. Posing this counterfactual question is a means of testing whether the loss for which damages are claimed is attributable to a matter which the defendant misrepresented or failed to report or whether it is a consequence of risks which the claimant would have taken upon itself even if the true position had been as the defendant represented it to be.

Properly applied, therefore, in Lord Hoffmann’s example of the mountaineer, the counterfactual test involves asking whether the mountaineer would still have been injured if his knee had in fact been sound as the doctor negligently advised that it was. In the valuers’ cases it involves asking what loss the lender would have suffered if the property mortgaged as security for the loan had in fact been worth what the valuer advised that it was worth. In an audit case the test involves asking whether the audited entity would still have suffered loss if its financial position had in fact been as the auditor represented it to be by its audit opinion. An affirmative answer to that question in the audit case indicates that the auditor is not responsible for the loss because it is a consequence of risks which the entity would have run even if the auditor’s opinion had been sound.

Even on a proper understanding of the counterfactual test, its application to the dividend example is problematic (as discussed by Fancourt J in BTI 2014 LLC v PricewaterhouseCoopers LLP [2020] PNLR 7, paras 112-121). I think that the solution to the problem requires recognising that, in applying the test, it is necessary to consider not simply how well off the claimant would be in a purely factual sense in the relevant counterfactual situation but whether the claimant would in that situation have suffered what the law regards as an injury. I will refer to a loss which the law regards as an injury as an “actionable loss”. A dividend paid out of distributable profits causes exactly the same diminution in the value of the company’s assets as one paid out of capital. In a purely factual sense, therefore, it could be said that the company has suffered the same loss in each case. However, it is only insofar as the payment of a dividend is unlawful (because it is not made from distributable profits) that the company suffers what the law regards as an injury for which a claim may be made. A dividend lawfully paid from distributable profits reduces the company’s assets in the same amount as a dividend unlawfully paid out of capital but does not cause any actionable loss.

Thus, in the example given at para 127 above where a company pays a dividend of £25m in reliance on the auditor’s negligent advice that there were distributable profits of £50m available, when on a true view there were none, the payment of the dividend causes an actionable loss to the company of £25m. Applying the counterfactual test, the relevant question is whether the payment of the dividend would have resulted in the same actionable loss if the auditor’s advice had been correct and the company’s financial position had been as represented in its audited accounts. In that situation the company would indeed have had profits of £50m out of which it could lawfully pay the dividend. Paying the dividend would therefore not have caused an actionable loss to the company. Accordingly, the loss incurred by paying the unlawful dividend is not excluded by applying the counterfactual test.

This example shows the importance of keeping in mind the underlying causal question which the counterfactual question is being used to test. It is not necessary to apply counterfactual reasoning in order to recognise that loss resulting from the payment of an unlawful dividend made in reliance on the auditor’s opinion is a loss caused by the matter which the auditor negligently failed to report - that is, the lack of any distributable profits out of which a dividend could lawfully be paid. If in such a situation applying the counterfactual test produces the “wrong” result, the most likely explanation is that the test has been incorrectly applied. The test is best seen, in my opinion, as a means of confirming or giving effect to the core principle that a professional adviser whose advice is a limited part of the material relevant to the client’s decision is responsible only for loss resulting from that decision which is a consequence of what makes the advice incorrect.

In SAAMCO (at p 220B) Lord Hoffmann said that to recover damages for negligent advice a claimant has to satisfy two separate requirements: “first, to prove that he has suffered loss, and, secondly, to establish that the loss fell within the scope of the duty he was owed.” In Nykredit (at p 1631D-1632A) Lord Nicholls, with the agreement of the other members of the appellate committee, made clear that the first of these steps involves comparing the claimant’s actual financial position with what that position would have been if the defendant had performed its duty. Lord Nicholls described this as “the basic comparison”; and in Platform Home Loans (at p 207E) Lord Hobhouse referred to the loss ascertained by making this comparison as the claimant’s “basic loss”. The second step then involves determining what part of this basic loss was within the scope of the defendant’s duty: see also Hughes-Holland, para 32. That requires, first of all, identifying the scope of the defendant’s duty.

I agree with the submission made by Mr Salzedo that it does not in principle matter in what order the claimant’s basic loss and the scope of the duty owed by the defendant are considered; it is, however, necessary to consider both before it is possible to determine whether all or part of the basic loss fell within the scope of the duty.

I will address these questions as they arise in the present case in the following order:

What basic loss did the society suffer as a result of Grant Thornton’s negligent advice?

What was the scope of Grant Thornton’s duty?

Did all or part of the society’s basic loss fall within the scope of Grant Thornton’s duty?

On this appeal there is no longer any dispute about the society’s basic loss. The starting point is the judge’s finding that, but for the negligent advice of Grant Thornton, the society would not have incurred the costs paid to close out the swaps in 2013 (see para 64 above). That is because, as the judge found, if Grant Thornton had given correct advice about hedge accounting from the outset, the society would not have entered into any more long-term interest rate swaps after April 2006 and would have closed out the swaps already taken out (which it could have done at that stage without loss). In that event the costs of £32.7m paid to close out the swaps in 2013 would not have been incurred.

It is also common ground that, if advised that its accounts could not properly be prepared using hedge accounting, the society would not have maintained its book of lifetime mortgages. It has accordingly been rightly accepted by the society that, if it recovers damages for losses suffered in pursuing its hedging strategy in reliance on Grant Thornton’s negligent advice, it must give credit for the gains made on the books of lifetime mortgages. These were £3.5m in the case of the UK mortgages and £2.46m (after taking account of administrative costs) in the case of the Spanish mortgages (see para 58 above), making £5.96m in total.

It follows that (in round numbers) the society’s basic loss suffered as a result of Grant Thornton’s negligent advice is £26.7m (£32.7m less £5.96m).

In accordance with the normal role of an auditor, Grant Thornton owed a duty to advise the society with reasonable care and skill whether its accounts had been properly prepared and gave a true and fair view of the society’s financial position. That included a duty to advise the society whether the lifetime mortgages and interest rate swaps entered into by the society were correctly accounted for in a manner which complied with applicable accounting rules and standards (including IAS 39, which governed hedge accounting). As with any statutory audit, the purpose was to provide the society and its members with accurate information on which they could rely in conducting its business of borrowing and lending money.

Equally, as noted earlier, it is no part of an auditor’s duty to give advice about how, in the light of such accounting information, the audited entity should conduct its business. Thus, it is agreed between the parties on this appeal that, in accordance with the judge’s findings, Grant Thornton did not advise the society whether it should enter into lifetime mortgages or interest rate swaps or about which particular mortgages or swaps to enter into. In particular, Grant Thornton gave no advice about the desirability of entering into swaps with 50-year terms. The judge found that the decision by the society’s board of directors to do so was based on a calculation that it was cheaper to enter into 50-year rather than 20-year swaps so long as LIBOR did not fall below a certain level and a commercial judgment (which proved to be misplaced) that LIBOR was unlikely to fall below that level. It was not part of Grant Thornton’s responsibility as auditor to advise the society about the merits of that business decision.

The advice which Grant Thornton gave in April 2006 that the society was entitled to apply the hedge accounting rules in IAS 39 to the lifetime mortgages and swaps (and to substitute a new mortgage as the item hedged by a swap when the original mortgage was redeemed) was similarly limited to the accounting treatment of those instruments. It is common ground that this advice was no different in nature (or relevant content) from the advice subsequently given about the use of hedge accounting in auditing the accounts. Again, Grant Thornton was not asked to and did not advise the society about the commercial wisdom of its intended hedging strategy or about whether it would be sensible or desirable to enter into interest rate swaps (and, if so, in what amounts and of what duration). Nor has it been suggested that it was any part of Grant Thornton’s remit, then or in its later audits, to assess or advise the society whether it had sufficient capital to carry on its business strategy while also meeting regulatory requirements, with or without the ability to use hedge accounting.

Accordingly, so far as relevant in this case, the purpose of Grant Thornton’s duty of care was solely to ensure that the society had accurate advice about the proper accounting treatment of the mortgages and swaps on which it could rely in taking commercial decisions, including the decisions which the society took to enter into long-term interest rate swaps as a hedge against changes in the fair value of its mortgage loans. It is apparent that the ability to use hedge accounting was, as Grant Thornton knew, critical to the adoption of the society’s business strategy. At the same time, Grant Thornton’s advice on hedge accounting, important as it was, was only part of the material relevant to the society’s decisions to enter into (and continue to hold) swaps and lifetime mortgages and regarding what particular swaps (of what duration) and mortgages to acquire. Those decisions were also based on other, commercial considerations concerning the costs, risks and benefits of its lending strategy which it was not Grant Thornton’s responsibility to identify or assess. It follows that, in accordance with the principles discussed at paras 90-94 and 115 above, Grant Thornton is not liable for all foreseeable consequences of the society’s reliance on its negligent advice but only for those which resulted from matters which made its advice wrong. In particular, Grant Thornton is not liable for losses which the society would have suffered as a result of entering into and retaining swaps and mortgages between 2006 and 2013 even if the accounting treatment of them adopted in the society’s accounts during that period had been correct, as Grant Thornton negligently advised that it was.

On the facts found and agreed for the purposes of this appeal, there were two relevant matters concerning the accounting treatment of the swaps and mortgages of which Grant Thornton negligently failed to inform the society and which made its advice given in April 2006 and in its subsequent audit opinions wrong. The first was that the characteristics of the swaps and lifetime mortgages did not meet the conditions required by IAS 39 to qualify for hedge accounting. In particular, there was no reasonable basis for expecting the hedges of the mortgages by the swaps to be “highly effective”. As mentioned earlier, one major reason for this was that there was no correlation (let alone match) between the duration of the interest rate swaps and the expected duration of the lifetime mortgages. Furthermore, the assumption made to get them to match - that, as and when mortgages were redeemed, new mortgages at the same fixed rate of interest could be substituted - was both unjustified and inconsistent with the hedge accounting rules.

The second matter, which is a corollary of the first, is that the society’s accounts were materially misstated. For example, the 2008 accounts showed a “loss on derivatives” of £26.364m. This represented a dramatic fall in the fair value of the interest rate swaps compared with their value at the end of 2007. However, this loss was offset in the accounts by a “gain on hedged items” of £26.798m, meaning that the value of the mortgages, as reported, had been increased by this amount compared with the value reported at the end of 2007. The adjustment to the carrying value of the mortgages was calculated by (unjustifiably) assuming that the mortgage book would continue to generate payments of interest at the existing fixed rate for the entire future period of the swaps (which would not mature until 2056 to 2058) and discounting this projected income stream to its net present value. It is accepted by Grant Thornton that this approach was flawed. It was flawed because (among other reasons) it was illegitimate to assume that, as and when mortgages were redeemed, new mortgages at the same fixed rate of interest could be substituted for them as items hedged by the swaps.

The impact of these accounting errors on the society’s reported financial position can be seen from what subsequently happened. When the 2011 accounts were restated after it was realised that they had previously been misstated, the net assets were reduced from £38.4m to £9.7m - a fall of £28.7m. This was the result of valuing the mortgages on the basis of their amortised cost (as required by the applicable accounting standards) instead of adjusting their value to offset the fair value of the swaps (on the mistaken view that hedge accounting could be used).

The question then is whether there was a causal connection between these matters and the society’s basic loss. It is plain that there was. The fact that there was no effective - let alone “highly effective” - hedging relationship between the fair value of the swaps and the impact of movements in interest rates on the value of the mortgages, as Grant Thornton had negligently advised that there was, had the result that, when the swaps were terminated, the MTM value of the swaps was not offset by a corresponding adjustment to the value of the mortgages.

On behalf of Grant Thornton Mr Salzedo submitted that the “value gap” between the swaps and the mortgages was not attributable to the inaccuracy of Grant Thornton’s advice because the society’s management was well aware that the duration of the swaps greatly exceeded the likely duration of the mortgages and therefore knew that in commercial terms the hedges were not effective despite what the accounts stated. He argued that the falsity of the accounts cannot in these circumstances be said to have been a cause of the society’s loss as the management was not misled by the accounts in their understanding of the society’s true financial position.

This argument, however, makes the mistake of focusing on the causal connection between Grant Thornton’s advice and the society’s loss instead of on the question whether the loss was causally connected to matters on which Grant Thornton was responsible for giving advice and which made the advice given incorrect. As discussed earlier (at paras 95-99 above), in considering whether loss suffered by the claimant was within the scope of the defendant’s duty, it is the latter question which is relevant. Whether and, if so, how the society’s management relied on the correctness of the accounts is relevant to the prior question whether the negligence of Grant Thornton caused the society to suffer any, and if so what, basic loss. But it is not relevant to the question now being considered of whether the basic loss which the society has established flowed from any of the matters of which Grant Thornton negligently failed to inform the society and which made its advice incorrect - one of which was the “value gap” which the accounts concealed. There is no doubt, as indeed Mr Salzedo emphasised in his submissions, that this value gap was a cause of the society’s loss in that, had there been no such gap, the costs incurred to close out the swaps would have been offset by a corresponding adjustment to the value of the lifetime mortgage books. It is immaterial in this regard whether or not the society’s managers were aware of the value gap.

In order to determine that the basic loss suffered by the society was causally related to the subject matter of Grant Thornton’s advice, there is no need to apply a counterfactual test; but equally there is no difficulty in doing so. If it had been true that, as Grant Thornton advised, the society’s financial position was accurately stated in its accounts, the hedges arranged by the society would have been “highly effective” (as defined in IAS 39 at AG105). For a hedging relationship to be classified as “highly effective”, there must be an actual and expected close (although not necessarily exact) match between changes in the fair value of the hedged item attributable to the hedged risk and changes in the fair value of the hedging instrument. The results reported in the society’s accounts in the years 2006 to 2011 showed a small difference each year between the reported movement in the fair value of the swaps compared with the adjustment to the carrying value of the mortgages. This varied between 0.8% (in 2011) and 13.5% (in 2008). Sometimes the difference represented a small reported net gain and sometimes a small reported net loss. It would be consistent with this variability and a scenario in which the hedge accounting rules had been correctly applied for the difference between the MTM value of the swaps and the adjustment to the value of the mortgages required to reflect interest rate risk at the time when the swaps were closed out to have been either a small negative or a small positive figure. However, there is no reason to expect that the difference is any more likely to have been to the detriment rather than to the benefit of the society. In these circumstances, the appropriate assumption to make for the purpose of assessing damages is that the value of the lifetime mortgages would have been higher than it was by an amount equal and opposite to the fair value of the swaps.

It follows that the full cost of closing out the swaps (leaving aside the transaction costs) is attributable to a risk (of the absence of an effective hedging relationship between the swaps and the mortgages) which Grant Thornton owed a duty of care to protect the society against. It therefore fell within the scope of Grant Thornton’s duty.

Unfortunately for the efficient resolution of these proceedings, this issue was confused by the way in which the society’s case has principally been argued. The society sought to argue that its basic loss was within the scope of Grant Thornton’s duty on the ground that this loss resulted from having to break the swaps before term and that the risk of that happening was a risk which Grant Thornton owed a duty of care to protect the society against. Counsel for the society submitted that the purpose of Grant Thornton’s advice was to avoid or mitigate the volatility from changes in the MTM value of the swaps to which the society’s balance sheet would be exposed unless hedge accounting could be used. Grant Thornton must have appreciated that the consequences of such volatility would or might well include demands for increased regulatory capital which hedge accounting was supposed to avoid and lead to the society having to break the swaps during their terms and thereby suffer loss. It was submitted that in these circumstances such loss was within the scope of Grant Thornton’s duty. As Teare J summarised the society’s argument at para 172 of his judgment:

This way of putting the society’s case should, in my view, be rejected for two main reasons. First, on the facts found by the judge, the society did not suffer loss from having to break the swaps before term. Its loss was caused by entering into swaps with 50-year terms and retaining them for several years (as a result of a combination of its own imprudence and Grant Thornton’s negligent advice). The evidence did not show that the loss thereby suffered would have been avoided or reduced if the society had not needed to break the swaps. To the contrary, the evidence indicated that the society’s loss would have been even greater if the swaps had not been broken. Second, for reasons which I will develop, the risks that the society’s balance sheet would become vulnerable to volatility and that the society might, in consequence, have to break the swaps prematurely were in any case not risks which Grant Thornton owed a duty of care to protect the society against.

On the first point, it is important not to confuse two propositions. It is no longer in dispute that, if Grant Thornton had given correct, non-negligent advice in 2006 that hedge accounting could not be used, the society would not have entered into long-term interest rate swaps (and would have closed out those already entered into) and hence would not have incurred the costs that were incurred when the swaps were closed out in 2013. That was how the society pleaded its case, and how it proved at the trial, that it suffered loss which was factually caused by Grant Thornton’s negligent advice. As discussed above, this “basic loss” was calculated by comparing its actual financial position with what its financial position would have been if it had not entered into any swaps (and had closed out those already entered into) after April 2006. The difference is the amount paid to break the swaps in 2013. It is this basic loss for which (after giving credit for the gains made on the mortgage books) the society is entitled to recover damages if and to the extent that Grant Thornton had a duty of care to protect the society against the occurrence of this loss.

It is a different proposition to assert that the society suffered loss from having to break the swaps before term. This second proposition involves a different comparison from the “basic comparison” used to calculate the society’s “basic loss”. It involves comparing the society’s financial position after paying the costs of breaking the swaps in 2013 with what its financial position would have been if, instead of breaking the swaps in 2013, it had continued to hold them until the end of their terms. In principle, it would have been possible for the society to show that it suffered a loss by having to break the swaps in 2013 as opposed to of holding them until the end of their terms as the society had planned to do. However, on the facts found by the judge the society failed to do so.

Leaving aside the transaction fees (of approximately £0.3m) which are no longer in dispute, the cost which the society incurred to break the swaps in June 2013 (of £32.4m) was the MTM value of the swaps at that time. That was the market value of the society’s obligation to make periodic payments at a fixed rate of interest, when netted off against the value of its right to receive such payments under the swaps at a variable rate of interest, for the remaining terms of the swaps. To show that it suffered a loss by having to break the swaps, the society would accordingly have had to prove that, had it continued to hold the swaps until the end of their terms, it would have made net payments under the swaps over this period which were less than the market forecast of such payments reflected in the MTM value of the swaps in June 2013. Developments between 2013 and the date of the trial provided no support for such a suggestion since, when the trial took place, the swaps were even further “out of the money” and their negative MTM value had increased from £32.4m to £56m. Furthermore, the judge was plainly justified in finding that, given the impossibility of predicting with any degree of reliability what will happen to financial markets, interest rates, the fortunes of the society and the views of the regulator over the next 30 or more years until the last swap would have expired in 2052, the society could not prove that on the balance of probability it would be financially better off if it had held the swaps to maturity (see para 66(iii) above).

On this appeal the society has sought to surmount this difficulty by submitting that the burden of proof is on Grant Thornton to establish that, if the swaps had been held to term, the society would have been obliged to make net payments under the swaps in an amount which was as great as their MTM value when they were terminated in 2013. The question of the burden of proof was considered by this court in Hughes-Holland, at para 53, where the view was taken that the burden is on the claimant to prove both that it has suffered loss and that the loss fell within the scope of the defendant’s duty. That conclusion was not necessary to the court’s decision in Hughes-Holland, as on the facts of that case it was found that the loss suffered by Mr Gabriel was not a loss which the defendant solicitors owed a duty of care to protect him against, irrespective of the incidence of the burden of proof (see para 19 of the judgment). Equally, I do not think it necessary to revisit the question in the present case. If it were necessary to decide whether (and, if so, by how much) the cost to the society would have been lower than the £32.4m (plus transaction fees) paid to terminate the swaps in June 2013 if the swaps had been held for their full terms, there was no better evidence than the MTM value of the swaps at the time of the trial. This represented the best estimate, as reflected in the market price, of what the cost to the society of holding the swaps until maturity would have been. Given the uncertainties involved, that evidence did not prove that this cost would on the balance of probability be £56m or even a sum close to that figure. But, as the best estimate, it signified that £56m was the mid-point in the range of potential outcomes such that the chance that the overall cost to the society of holding the swaps to the end of their terms would be greater than £56m was equal to the chance that it would be less than this amount. That was a sufficient basis on which to conclude that, on the balance of probability, the society would have been financially worse off if it had continued to hold the swaps rather than breaking them in 2013.

For these reasons, which are in substance the same as those given by Hamblen LJ in the Court of Appeal, the society’s assertion that it suffered loss as a result of having to break the swaps prematurely is one which cannot be sustained on the facts found in this case.

The second reason why the society’s argument should in my view be rejected is that, even if it had been shown that, on the facts, the society suffered a loss from having to break the swaps, that would not be a loss which Grant Thornton owed a duty of care in law to protect the society against. The suggestion that Grant Thornton owed a duty to protect the society against the risk of volatility to which its balance sheet would be exposed if it was unable to use hedge accounting (and the consequences of that risk materialising) treats Grant Thornton as if it had given a warranty or guarantee that the society was and would be entitled to use hedge accounting over the lifetime of the swaps. That, however, was not the nature of Grant Thornton’s duty. Grant Thornton did not undertake any responsibility to ensure that the society was or would be entitled to use hedge accounting - still less to ensure that, by using hedge accounting, the society would have (either at the outset or in future) sufficient capital to meet regulatory requirements and also carry on its mortgage lending strategy. Grant Thornton’s duty was simply the ordinary one owed by a professional adviser to take care to give accurate advice - in this case accurate advice about whether or not hedge accounting could properly be used.

An adviser who, in breach of such a duty, gives incorrect advice (in this case, that hedge accounting could properly be used in preparing the society’s accounts) is not liable, on any view, to pay damages calculated to put the society in the same financial position as if it had been entitled to use hedge accounting. That is the measure of damages applicable to a claim for breach of a contractual warranty. In accordance with the principles established by SAAMCO and Hughes-Holland and not disputed on this appeal, such an adviser is not even liable for all the foreseeable adverse consequences of decisions taken in reliance on its incorrect advice. It is liable only for consequences of matters which made its advice incorrect - in this case the fact that there was in truth no effective hedging relationship between the swaps and mortgages. It was that risk which Grant Thornton owed a duty of care to protect the society against.

The society’s argument on this point (summarised at para 151 above) seems to me to confuse the purpose of Grant Thornton’s advice and duty of care with the purpose for which the society intended to rely on Grant Thornton’s advice. An essential part of what SAAMCO and Hughes-Holland decided is that the defendant’s awareness that the claimant is relying on its advice for a particular purpose does not itself make the defendant responsible for the claimant’s failure to achieve that purpose, if the advice proves to have been negligent and wrong. It bears repeating that, as Lord Sumption put it in Hughes-Holland at para 41:

Expressed in terms of purpose, the purpose of Grant Thornton’s advice and duty of care, objectively considered, was to ensure that the society was provided with accurate advice about the proper accounting treatment of the swaps and mortgages. Grant Thornton’s liability for the foreseeable consequences of giving incorrect and negligent advice is correspondingly limited by the scope of duty principle to such consequences as flowed from matters which made that advice incorrect.

A loss which resulted from having to break the swaps before the end of their terms would not even have been a loss caused by adopting an improper accounting treatment in reliance upon the accuracy of Grant Thornton’s advice. Rather, it would have been a loss caused by discontinuing an improper accounting treatment after Grant Thornton had realised its error and advised that (contrary to its earlier advice) the society had been preparing its accounts on an incorrect basis and was not entitled to use hedge accounting. The society has no right to complain that Grant Thornton did not continue to give it wrong advice after 2012. Nor has it any right to base a claim for damages on its inability to continue to pursue a strategy which was dependent on presenting its accounts in a way which was improper and gave a false view of its financial position. A claim for damages for loss suffered as a consequence of being compelled to break the swaps would therefore be bad in law.

It seems to me that the society’s representatives may have been led into these errors, at least in part, by the misunderstanding discussed earlier of what is meant by asking - to test whether the claimant’s loss falls within the scope of the defendant’s duty - whether the loss would have occurred if the defendant’s advice had been correct. As can be seen, for example, in the passage in Teare J’s judgment summarising their argument quoted at para 151 above, counsel for the society sought to satisfy the counterfactual test in the present case by arguing that, had the advice given by Grant Thornton that hedge accounting could be used been correct, the society’s balance sheet would not have been vulnerable to volatility, in which case there would have been no need to break the swaps and the costs of doing so would not have been incurred. Therefore, it is suggested, the society can demonstrate that it suffered a loss which was within the scope of Grant Thornton’s duty by showing that the loss would not have been suffered if Grant Thornton’s advice had been correct.

Leaving aside the difficulties I have already mentioned that the society was in fact unable to show that it suffered any loss from having to break the swaps prematurely, that Grant Thornton owed no duty of care to protect the society against such a loss, and that a loss of this nature would not be the basic loss which the society has pleaded and succeeded in proving at the trial, it can also be seen that the society’s argument involves the same misunderstanding of the counterfactual test discussed at paras 126-129 above. That is because it treats the relevant question as being whether the society would have acted differently - and, if so, what the outcome would have been - if Grant Thornton had given the advice which it in fact gave (that the society’s accounts had been properly prepared in compliance with the hedge accounting rules) but, instead of being wrong, this advice is supposed to have been correct advice to give. The argument is that, if Grant Thornton’s advice that the accounts had been properly prepared had been correct advice to give, then the society would not have had to break the swaps and would instead have acted differently by holding the swaps to term. (It is then contended that, if the society had held the swaps to term, the cost of doing so would have been nil or at any rate less than the amount paid to close out the swaps in 2013.)

It makes no sense, however, and serves no rational purpose to suppose that it was correct for Grant Thornton to advise that the society’s accounts had been properly prepared in compliance with the hedge accounting rules, even though this was undoubtedly not the case because there was a major mismatch between the expected durations and fair values of the swaps and mortgages. No rationale has been suggested, and I can see no possible reason, for postulating such a counterfactual situation. As discussed earlier, the correct counterfactual question is not whether the claimant would have acted in the same way if (on the actual facts) the advice given by the defendant is supposed paradoxically to have been correct advice to give. It is whether, if the defendant’s advice had been correct in the sense that the facts had been as the defendant represented them to be, the actions which were in fact taken by the claimant would have resulted in the same (actionable) loss as the claimant in fact suffered. Thus, in the present case the relevant question is whether, if the society’s financial position had been as Grant Thornton represented it to be in its opinions on the accounts, the same loss would have occurred on breaking the swaps (and selling the mortgages) as was in fact suffered. As discussed earlier, the reason for asking this question is to test whether the loss suffered by the society was a consequence of a matter which made the defendant’s advice wrong. The society’s formulation of the counterfactual question misses this point.

The way in which the society’s case was principally advanced obscured in the courts below the correct way of analysing its claim. It all but did so again on this appeal. As I have emphasised, the society’s pleaded case is that it is entitled to damages for loss suffered as a result of entering into the swaps in reliance on Grant Thornton’s negligent advice. However, in its written argument on this appeal the society said, confusingly, that it is “not complaining about the purchase of the swaps, but about the fact that it has had to terminate the swaps …”. It also said that a claim based on entering into the swaps as a result of Grant Thornton’s negligent advice “is not the claim the society makes in these proceedings, which is only for the costs of being required to terminate a business that it would otherwise have continued to pursue …” (emphasis in the original).

In oral argument Ms Rebecca Sabben-Clare QC continued to focus on this version of the society’s case. She did, however, also submit that, if Grant Thornton’s advice had been correct, there would have been an effective hedging relationship between the swaps and the mortgages and hence no gap between the negative value of the swaps and the positive value of the mortgages - which would have cancelled each other out. Grant Thornton had a fair opportunity to meet this argument, as Mr Salzedo understood the society’s case to be put in both ways and, in his phrase, to “seesaw” between them. I have addressed the objection made by Mr Salzedo on behalf of Grant Thornton to this second way in which the society’s case was put at para 147 above.

For the reasons given at paras 143-149 above, this version of the society’s case is in my view well-founded. Had there been an effective hedging relationship between the swaps and the mortgages as Grant Thornton advised that there was, the society would not have suffered the loss which it suffered as a result of entering into long-term swaps - as in such circumstances, whenever the swaps were closed out, the fair value of the swaps would have been offset by a corresponding difference in the fair value of the mortgages. Thus, the cost of closing out the swaps which constituted the society’s basic loss (before giving credit for the fair value of the mortgages) resulted from a risk (of the lack of an effective hedging relationship between the swaps and the mortgages) to which, if Grant Thornton’s advice had been correct, the society would not have been exposed and which Grant Thornton owed a duty of care to protect the society against. Accordingly, the loss was within the scope of Grant Thornton’s duty.

Although I did not understand counsel for Grant Thornton to place any reliance on them, I will also address briefly the judge’s reasons, not based on either party’s submissions at the trial, for holding that the society’s losses were not within the scope of Grant Thornton’s duty. The judge treated this question as requiring the court to make a broad evaluative judgment, based on “stand[ing] back and view[ing] the matter in the round”, as to whether Grant Thornton had assumed responsibility for the losses (see para 179 of his judgment). In my opinion, counsel for Grant Thornton were right not support this approach.

The judge derived his approach from a statement made by Lord Sumption in Hughes-Holland at para 44, after contrasting a case involving a valuer or conveyancer with one involving an investment adviser:

The judge collapsed these two stages of the analysis into one. He also appears to have understood the question whether the loss flowed from the particular features which made the defendant’s conduct wrongful as merely one element to be weighed in an overall evaluation rather than as conclusive of whether the loss was within the scope of the defendant’s duty (in a case where the claimant has established that it relied on the defendant’s negligent advice for an intended purpose and suffered loss as a result). In addition, in reaching his ultimate conclusion (at para 179 of the judgment), the judge adopted the society’s misconception of its loss as caused by having to break the swaps when it was appreciated that Grant Thornton’s advice was wrong. Thus, the judge attached particular weight to the fact that the same loss would have been sustained if the counterparties had decided to terminate the swaps early, as could have happened at any time. That would have been a valid objection to the society’s case (in addition to all those discussed above) if its loss is conceived as the consequence of breaking the swaps before the end of their terms. On a correct analysis, however, it is not a relevant consideration. If the society’s financial position had been as Grant Thornton represented it to be, it would not have mattered whether the swaps were terminated by the society or by the counterparties. In either case the society would not have suffered the loss that it did in fact suffer because the sums paid to the counterparties when the swaps were closed out would have been offset by a corresponding increase in the fair value of the mortgages.

As for the judge’s comment, endorsed by the Court of Appeal (at para 99 of the judgment), that it would be “a striking conclusion to reach” that an accountant who advises a client as to the manner in which its business activities may be treated in its accounts is legally responsible for the financial consequences of those business activities, I cannot accept that there is anything inherently unlikely or surprising about such a conclusion. As the majority of the Supreme Court of Canada observed in Livent (at para 87):

If, as I consider, the society’s basic loss was within the scope of Grant Thornton’s duty, Grant Thornton submits that the decision of the judge and the Court of Appeal should still be upheld on the alternative ground that the judge was wrong to conclude that in law the loss was caused by Grant Thornton’s negligence. I find it difficult to see that there is any room for such an argument. In order to decide that the claimant’s loss was within the scope of the defendant’s duty, the court must be satisfied that the loss was caused by the particular matters which made the defendant’s advice incorrect and not by other matters unrelated to the subject matter of the defendant’s negligence. That seems to me to be sufficient to demonstrate that the defendant’s conduct was an effective cause of the loss.

In any event, I am unable to accept that on the facts the judge was bound to conclude that the society’s decision to enter into swaps with terms far longer than the likely duration of the mortgages, combined with the way that interest rates turned out, were the only effective causes of its loss. In my view, he was fully entitled to conclude that an equally effective cause was Grant Thornton’s negligent professional advice, maintained over a period of some seven years, that it was permissible to use hedge accounting and prepare accounts which showed the swaps to be a highly effective hedge for the lifetime mortgages, thereby hiding the mismatch between the values of the swaps and mortgages and the society’s inadequate regulatory capital. The society’s own negligence which contributed to this state of affairs was properly reflected in the reduction of 50% which the judge thought it just and equitable to make to any damages awarded.

For all these reasons, I conclude that the judge and the Court of Appeal were wrong to hold that the loss sustained by the society as a result of entering into long term interest rate swaps in reliance on Grant Thornton’s negligent advice was not within the scope of Grant Thornton’s duty. They should have concluded that it was a loss from which Grant Thornton owed a duty of care to protect the society. The loss was caused by a matter - the lack of an effective hedging relationship between the swaps and the lifetime mortgages which they were supposed to hedge - which Grant Thornton negligently failed to appreciate and report to the society and which made its advice wrong. By the same token, if Grant Thornton’s advice had been correct and there had been an effective hedging relationship between the swaps and the mortgages, as Grant Thornton advised that there was, the loss would not have occurred.

It follows that, after giving credit for the value of the mortgages and the 50% reduction for its contributory negligence, the society is entitled to recover damages, in addition to the amount awarded by the judge, of some £13.4m (50% of £26.7m). The exact figure can no doubt be agreed.

I agree that this appeal should be allowed. In preparing this judgment, I have had the benefit of reading the judgment of Lord Leggatt. I agree with most of his central reasoning. But, with respect, I do not agree with all aspects of his analysis of South Australia Asset Management Corpn v York Montague Ltd [1997] AC 191 (“SAAMCO”). In particular, my view is that the “SAAMCO principle” (which can also be referred to as the “scope of duty principle”) can most easily be understood without regarding it as based on causation. Subsequent to preparing this judgment, I have also had the benefit of reading the joint judgment of Lord Hodge and Lord Sales; and, as I explain in more detail at para 212, their essential reasoning and mine are closely aligned. The purpose of this judgment is to explain in my own words how I understand SAAMCO and how it applies in this case.

This is a case involving an auditor’s negligence. The central question at issue is how SAAMCO applies to the facts. It is the first case to reach the highest court in which it has had to be decided how SAAMCO applies to an auditor’s negligence. It is being heard in parallel with Khan v Meadows [2021] UKSC 21 which raises the question of how SAAMCO applies to a claim for a doctor’s negligence. As I understand it, a seven-person panel was convened in each case so as to be able to deal fully with any submission by the claimants that the Supreme Court should depart from SAAMCO. In the event, no such submission has been made in either case. Even if such a submission had been made, the Court would have been reluctant to accede to it. While it is hard to deny that SAAMCO was, and remains, controversial, it has been confirmed and applied in many cases including by the Supreme Court most recently in Hughes-Holland v BPE Solicitors [2017] UKSC 21; [2018] AC 599 (“Hughes-Holland”). The doctrine of precedent - and the stability which it engenders - dictates that departures from a settled line of House of Lords or Supreme Court authority, by use of Practice Statement (Judicial Precedent) [1966] 1 WLR 1234, should be rare.

The SAAMCO principle is best understood as a principle which focuses on the scope of the defendant’s duty of care (whether in contract or tort). In almost all past cases, applying SAAMCO, the context has involved a defendant providing professional services, through advice or information, to the claimant. It is unnecessary for the purposes of this case (and Khan v Meadows) to consider whether - and, if so, when and how - SAAMCO may apply outside that context. In the context with which we are dealing, one can say that the SAAMCO principle is concerned to determine whether factually caused loss is within the scope of the professional’s duty of care to the claimant. The loss in question will almost always be pure economic loss (although it could be property damage or personal injury, as in Lord Hoffmann’s famous mountaineering example in SAAMCO). The SAAMCO principle, focused on the scope of the duty of care, is underpinned by a policy of seeking to ensure that a professional’s liability for negligence reflects a fair and reasonable allocation of the risk of the loss that has occurred as between the parties. The SAAMCO principle is generally regarded as imposing a limit on liability different from the restrictions of remoteness and legal causation (the latter can alternatively be labelled “intervening cause”) although whether that is so may depend on what one regards as determining remoteness and legal causation. So, for example, if one regards remoteness as dealing purely with the foreseeability or contemplation of the type of loss, one can readily treat the SAAMCO principle as a separate limiting principle. If, on the other hand, one regards remoteness as embracing an “assumption of responsibility” test (as in Transfield Shipping Inc v Mercator Shipping Inc (The Achilleas) [2008] UKHL 48; [2009] AC 61, in relation to remoteness in contract), remoteness might be thought to be wide enough to embrace the SAAMCO principle.

In this case, an auditor, Grant Thornton UK LLP, the defendant and respondent (who I shall refer to as the defendant or “Grant Thornton”), negligently advised the Manchester Building Society, the claimant and appellant (who I shall refer to as the claimant or “the society”), that it could apply to its accounts a form of accounting known as hedge accounting. From April 2006, the society relied on this advice in formulating its business plan by which it entered into interest rate swap agreements as a hedge against interest rate changes affecting its central lending and mortgage business; and it continued to rely on that advice, and on the approval of its accounts by Grant Thornton, even though the accounts, using hedge accounting, served to obscure the true financial position of the society. After several years, Grant Thornton’s negligence came to light, and the society realised that it ought not to have been applying hedge accounting and could not continue to do so. With the encouragement of the regulator, it therefore closed out the swaps incurring break costs of £32.7m in June 2013 comprising payments made to the counterparties based on the fair market value of the swaps at the time that they were closed out (plus “transaction costs”). The society seeks damages for the main loss of £32.7m. We are solely concerned with that head of loss and not with the other heads of loss claimed, some of which were rejected by Teare J at trial and some of which were accepted by him: [2018] EWHC 963 (Comm); [2018] PNLR 27. For example, there has been no appeal by Grant Thornton against Teare J’s decision, at para 211, that Grant Thornton is liable to pay the society its “transaction costs” of breaking the swaps, which he assessed at £285,460.

Teare J refused the claim for £32.7m by an application of SAAMCO. He thought it unnecessary to decide whether this was an advice or an information case. This was the distinction drawn by Lord Hoffmann (giving the leading speech in the SAAMCO case) but, as we shall see, recognised to be “descriptively inadequate” in Hughes-Holland, at para 39. Teare J held that, applying SAAMCO, Grant Thornton was not liable for the £32.7m because it had not assumed responsibility for that type of loss. Although not arising on his reasoning, he also expressed the view that, applying the counterfactual referred to in SAAMCO, the opposite result would have been reached because the society could prove that, if the information or advice had been correct, the same loss of £32.7m would not have been suffered as at June 2013. Teare J’s view was that, had SAAMCO not ruled out recovery, remoteness and legal causation would not have done so. The type of loss was in the contemplation of the defendant as not unlikely to result from its negligence and was therefore not too remote (para 201). And there was no break in the chain of causation so that the breach of duty was an effective legal cause of the loss (paras 140-149).

The Court of Appeal, with the leading judgment being given by Hamblen LJ, as he then was (with whom Males and Gloster LJJ agreed), dismissed the claimant’s appeal against that decision and agreed with Teare J that SAAMCO ruled out the claim: [2019] EWCA Civ 40; [2019] 1 WLR 4610. However, Hamblen LJ disagreed with Teare J as to how exactly SAAMCO here operated. In Hamblen LJ’s view, SAAMCO did not support the application of a free-standing assumption of responsibility approach. Rather SAAMCO ruled out damages for the £32.7m because this was an information, not an advice, case and the claimant could not satisfy the burden of proof on the claimant in relation to the counterfactual required by SAAMCO. The claimant now appeals to the Supreme Court.

For fuller details of the facts and the judgments below, I am grateful to rely on the judgment of Lord Leggatt. I shall also not repeat his description of the facts and decisions in the leading cases of SAAMCO and Hughes-Holland.

Although at some stages of the hearing before us, this may have been thought unclear, this appeal is about how SAAMCO applies. It is not about whether the claimant has suffered a loss factually caused by the defendant’s negligence. Teare J found that factual - ie but for - causation was here proved by the claimant. So had the defendant performed its contractual or tortious duty of care, Teare J found that the claimant would not have suffered the loss of having to pay the swap break cost of £32.7m in 2013 (see his judgment at para 139). To explain this further, that was a factually caused loss because Teare J found that, had the defendant not been negligent in April 2006, the claimant would not have entered into more long-term hedging swaps and would have closed out the swaps it had already entered into (see para 127). The claimant would therefore not have had to break those swaps, as it reasonably did, in 2013 and the cost of so doing was £32.7m.

There are two additional clarificatory points on this:

The question at issue is whether that factually caused loss falls within the scope of Grant Thornton’s duty of care, and is therefore recoverable, or falls outside it and is therefore irrecoverable.

In seeking to understand SAAMCO, I have found the following articles, essays, and case-notes especially helpful: Jane Stapleton, “Negligent Valuers and Falls in the Property Market” (1997) 113 LQR 1; Janet O’Sullivan, “Negligent Professional Advice and Market Movements” [1997] CLJ 19; John Wightman, “Negligent Valuations and a Drop in the Property Market: the Limits of the Expectation Loss Principle” (1998) 61 MLR 68; John Murdoch, “Negligent Valuers, Falling Markets and Risk Allocation” (2000) 8 Tort Law Rev 183; Edwin Peel, “SAAMCO Revisited” and Richard Butler, “SAAMCO in Practice” in Commercial Remedies (eds Andrew Burrows and Edwin Peel, 2003) pp 55-87; Hugh Evans, “Solicitors and the Scope of Duty in the Supreme Court” (2017) 33 PN 193; Desmond Ryan, “SAAMCO re-explored: BPE and the Law of Professional Negligence” (2018) 34 PN 71.

I have explained that the SAAMCO principle is concerned to determine whether factually caused loss is within the scope of the professional’s duty of care. The question as to the scope of the professional’s duty of care is a question of law underpinned by policy. But clearly it depends on a close analysis of the facts as to what the defendant has said and done and what the parties understood. As we shall see, the purpose of the advice or information, whether the cause of action for negligence is in tort or contract, is of particular importance in working out the scope of the duty.

Lord Hoffmann in SAAMCO relied on a counterfactual test as a guide to whether the loss falls within the scope of the duty of care. In order to apply this test, Lord Hoffmann reasoned that one first has to decide whether the case is one of advice or information. If it is an advice case, the counterfactual test does not apply and the scope of the duty of care extends to the recovery of all factually caused losses subject to the standard limitations of remoteness and legal causation. But if it is an information case, the scope of the duty of care is limited to the recovery of loss consequent on the information being wrong. That ensures the fair and reasonable allocation of risk as between the parties. Applying the counterfactual test, one asks, would the claimant still have suffered the same loss if the information had been true? If the answer is “yes”, the scope of the duty does not extend to the recovery of that loss. If the answer is “no”, the scope of the duty does extend to the recovery of that loss.

Lord Hoffmann explained the policy in play in the following way at p 214:

Also helpful in appreciating the underlying policy behind SAAMCO is the explanation given by Lord Nicholls in Nykredit Mortgage Bank Plc v Edward Erdman Group Ltd [1997] 1 WLR 1627, another negligent valuer case, which followed SAAMCO, at p 1631:

Therefore, both Lord Hoffmann and Lord Nicholls saw the restriction to “the consequences of the information being wrong” as being concerned to achieve a fair and reasonable allocation of the risk of the loss that has occurred as between the parties. It follows that the counterfactual test seeks to reflect that underlying policy. It serves as a guide - one may regard it as a useful cross-check in most cases - in working out whether the loss in question is within the scope of the duty of care.

It is important to clarify that the restriction to the consequences of the information being wrong cannot be explained as a matter of factual causation flowing from the nature of the wrong in question. This is not least because, as several commentators have correctly stressed (see, eg, Jane Stapleton, “Negligent Valuers and Falls in the Property Market” (1997) 113 LQR 1, 3), the wrong in the valuer cases was not the breach of a warranty that the information given was correct. The wrong was simply the failure to use reasonable care in supplying the valuation figure that was given. Had reasonable care been taken, the valuer would have given an accurate valuation within a range: but that is not the same, even in a rough and ready way, as saying that the valuation that was given should be treated as correct for the purposes of a causal enquiry based on the wrong. Rather than articulating an alternative causation justification, it is my view that the clearest explanation for why the defendant professional is liable only for the consequences of the information being wrong is that this rests on the policy that it would not be fair and reasonable (Lord Hoffmann, at p 214, also said that it would be “paradoxical”) to protect the claimant against the risk of loss that would have been incurred and borne even if the professional had been warranting (which it was not doing) that the information was correct. The risk of that loss is one that should be borne by the claimant.

In a recent thought-provoking contribution to the debate about SAAMCO, Jane Stapleton has also stressed the particular policy underpinnings of the SAAMCO principle. In Three Essays on Torts (2021) she writes at pp 97-98 that:

When, in general terms, is it helpful to apply the counterfactual test? Lord Hoffmann in SAAMCO saw it as ideally suited for assisting in delimiting the scope of duty in relation to the giving of information as opposed to advice. However, it has subsequently become clear, as recognised in Hughes-Holland, that the information and advice distinction is problematic as being too rigid. Lord Sumption said at para 39:

Having said that, it is not easy to find shorthand replacement terminology. What may be said is that the more limited the advice or information being provided - in the sense that the more the claimant has to decide on - the more appropriate the counterfactual test is likely to be. Certainly, it is important to recognise that there is a spectrum of professional liability for misstatements, or failures to inform, that runs from the giving of wide-ranging advice to the provision of specific information. Take the question, facing the courts in the valuer cases like SAAMCO, of whether the professional defendant was liable to its client for losses that included losses consequent on the sharp fall in the property market. Had the professional been an investment or pension adviser, advising the client as to which investments to make, it would be appropriate for the negligent professional to bear the risk of the client’s investment loss (including market falls). The scope of the duty would extend to such a loss because guarding against the risk of such a loss was the purpose of the client seeking professional advice rather than exercising his or her own judgment and was the purpose of the professional providing that advice. In contrast, as on the facts of SAAMCO, where the client was a lender seeking information as to whether a property was of sufficient value to provide adequate security for a loan, it would not be appropriate for the negligent professional to bear the risk of a loss by reason of a market fall (as opposed to the risk of loss consequent on the property not providing the relied upon cushion of security for the lender at the time of the valuation). Guarding against a market fall was not the purpose of the lender seeking, or of the professional providing, the valuation of the property.

That the advice and information categories are on a spectrum, with investment advice at one extreme and a valuer’s information (or, for example, a solicitor’s conveyancing information) at the other extreme, with many cases in between, was made clear by Lord Sumption in Hughes-Holland at para 44:

That the counterfactual test is of second-order importance in determining whether the loss in question falls within the scope of the duty of care owed by the professional is well-illustrated by the difficulties the courts have encountered in applying the SAAMCO counterfactual to cases on an auditor’s negligence. To illustrate this, I shall refer to just one recent case.

In BTI 2014 LLC v PricewaterhouseCoopers LLP [2019] EWHC 3034 (Ch); [2020] PNLR 7, a central claim was for a company’s loss in paying out a dividend to shareholders in reliance on an alleged negligent audit report. The defendant auditors sought reverse summary judgment or to have the claim struck out on the basis that, applying SAAMCO, that loss was irrecoverable. That application failed on the grounds that rejection of such a claim raised novel issues that were not suitable for summary determination without clear findings of fact. The reasoning on this issue of Fancourt J (at paras 108-121) is instructive and can be summarised as follows:

We therefore here see Fancourt J struggling to accommodate, on the one hand, his intuition, supported by past authorities and the general understanding of the law, that the scope of an auditor’s negligence does extend to the company’s loss in paying out dividends that would not have been paid but for the auditor’s negligence, and, on the other hand, the application of the SAAMCO counterfactual test which, on one approach, would reach the opposite result.

The way out of the conundrum is to recognise that the decision as to the scope of the duty of care focuses, especially, on the purpose of the advice or information and is underpinned by the policy of achieving a fair and reasonable allocation of the risk of the loss that has occurred; and that the counterfactual test is merely a cross-check on that decision which will not be helpful in all cases. The scope of an auditor’s duty of care plainly extends, in a standard case, to the loss of incorrectly paying out dividends in reliance on the auditor’s report because that is an obvious purpose of the company in having the audit report prepared. The purpose of an audit report for a company is for the shareholders to know that the accounts present a true picture of the financial state of the company so that, for example, any mismanagement by the directors can be exposed and decisions can be made as to whether it is appropriate to pay, and the value of, dividends to the shareholders. In the light of that purpose, it is fair and reasonable that the risk of loss by incorrectly paying out dividends should be borne by the auditor. The fact that one version of the counterfactual test gives the opposite result should give one pause for thought but should not override the decision on the scope of the duty which has been arrived at for reasons independent of the counterfactual test.

I am not suggesting that difficulties in applying the SAAMCO counterfactual are only encountered in relation to an auditor’s negligence. On the contrary, commentators have shown that there are problems in applying it in other areas (ie it reaches results that do not accord with one’s reasoned view as to what the correct allocation of risk should be) such as a solicitor’s negligence: see, eg, Hugh Evans, “Solicitors and the scope of duty in the Supreme Court” (2017) 33 PN 193.

The important point to stress, therefore, is that the decision as to whether loss falls within the scope of the professional’s duty of care is a question of law, with a particular emphasis on the purpose of the advice or information, that is underpinned by the policy of achieving a fair and reasonable allocation of the risk of the loss that has occurred as between the parties. Applying a counterfactual test can assist, whether one regards the case as one of advice or information, but such a test merely operates as a cross-check on one’s decision as to that allocation of risk. Moreover, there is some room for choice in the precise counterfactual test that is used.

We are now in a position to apply the SAAMCO principle to the facts of this case. It is not in dispute that Grant Thornton was negligent in advising that hedge accounting could be used. But very importantly, it is also not in dispute that, in the context of that advice, Grant Thornton negligently misstated the underlying financial position of the society in the specific respect that, in truth, there was no effective hedging relationship between the swaps and the mortgages (see Lord Leggatt’s judgment at para 168).

As we have seen in paras 184-185 above, had Grant Thornton performed its contractual or tortious duty of care properly, the society would not have suffered the overall factual loss of £26.7m. The question posed by the SAAMCO principle is whether the factually caused net loss (£26.7m) was within the scope of the auditor’s duty of care. That is a question of law, focusing especially on the purpose of the advice or information, and rests on the underlying policy as to the fair and reasonable allocation of the risk of loss as between the parties.

It was of critical importance to the society, in pursuing its business model, to know whether hedge accounting was acceptable or not. Grant Thornton advised the society that it was. The purpose of that advice was clear to Grant Thornton in that it knew that the society was explicitly relying on that advice in pursuing its business model. That in itself might, perhaps, not have been enough to reach the conclusion that the risks consequent on adopting that business model were appropriately borne by Grant Thornton. But the crucial additional factor that makes it clear that it was fair and reasonable for the risk of the loss to be borne by Grant Thornton was the negligent specific misrepresentation that there was an effective hedging relationship between the swaps and the mortgages. It was that specific misrepresentation, in the context of the advice on hedge accounting, that meant that the business model was pursued despite the society having insufficient regulatory capital. Clearly Grant Thornton knew that the purpose of that representation was to provide a true picture of the society’s financial position on which the society would rely in pursuing its business model. In my view, therefore, the society has established (the burden of proof being on the society, as claimant) that the loss was within the scope of Grant Thornton’s duty of care.

What about the SAAMCO counterfactual test? As I have been at pains to make clear, the application of a counterfactual test is not always helpful even if one is dealing with the provision of fairly specific information rather than wide-ranging advice. But what answer would be given by applying such a test? Both Teare J and the Court of Appeal in this case regarded the relevant counterfactual test (if applicable) as requiring the court to ask, would the same loss have been suffered had it been true that it was acceptable to use hedge accounting? Applying that counterfactual test, it is clear that the society would not have broken the swaps, and thereby suffered the break cost (of £32.7m) in June 2013. That counterfactual test might, therefore, be thought to support the view that the loss is recoverable. That was the view on the counterfactual test reached by Teare J and, because it contradicted his view as to Grant Thornton not having assumed responsibility for the break loss, he decided that the application of a counterfactual test was here unhelpful. However, the Court of Appeal thought that Teare J’s application of the counterfactual was incorrect because, if it had been true that it was acceptable to use hedge accounting, one would need to run the counterfactual through to the end of the terms of the long-term swaps. And as the best present evidence (taking the fair market value of the swaps) was that, at the very least, the society would have suffered the same loss (and the present evidence indicated a much greater loss) at the end of the swaps, the application of the counterfactual supported, rather than contradicted, the view that the loss was outside the scope of Grant Thornton’s duty of care.

One might baulk at that application of a counterfactual test by the Court of Appeal on the grounds that it is placing an inappropriately difficult burden of proof on the society. Counsel for the society, Rebecca Sabben-Clare QC, submitted that, by analogy to, for example, the contractual rule that it is for the defendant to show that the claimant’s proved reliance loss does not exceed the expectation loss (see Omak Maritime Ltd v Mamola Challenger Shipping Co (The Mamola Challenger) [2010 EWHC 2026 (Comm); [2011] Bus LR 212; [2011] 1 Lloyd’s Rep 47), once the society had shown that the loss would not have been suffered in 2013, the burden of proof in relation to an extension of the counterfactual to the end of the swaps should lie on Grant Thornton not the society. However, any argument for a modification of the burden of proof faces the challenge that, while not essential to the decision, in Hughes-Holland, para 53, Lord Sumption, with whom the other members of the Supreme Court agreed, took the view that the legal burden of proof in relation to SAAMCO was on the claimant not the defendant. In any event, one would have thought that the present market value is the best evidence of the value of the swaps at the end of their term, whichever party has the burden of proof.

However, in my view, in agreement with Lord Leggatt, that was not the most helpful counterfactual to apply. I have indicated in para 206 above that, on these facts, the crucial additional factor in here allocating the risk of loss was the specific misrepresentation that there was an effective hedging relationship between the swaps and the mortgages. In line with that, the most helpful counterfactual test is, with respect, not the one applied by Teare J or the Court of Appeal. Rather the question one should be asking is, would the same loss have been suffered had it been true that there was an effective hedging relationship between the swaps and the mortgages? If that had been true, the break cost (of £32.7m) would clearly not have been suffered (because, as gains and losses from movements in interest rates would roughly match, there would have been no significant break cost). It follows that this version of the counterfactual test supports the view that the net loss of £26.7m is recoverable as falling within the scope of Grant Thornton’s duty of care.

This analysis neatly illustrates that there is flexibility in the application of a counterfactual test. It performs a useful function as a cross-check in most cases but the crucial decision as to the scope of the duty of care focuses especially on the purpose of the advice or information and rests on the underlying policy as to the fair and reasonable allocation of the risk of loss as between the parties. On these facts, an examination of the purpose of the auditor’s advice and the specific misrepresentation that there was an effective hedging relationship between the swaps and the mortgages make clear that, subject to any contributory negligence by the society, it is fair and reasonable for the risk of the break cost to be allocated to Grant Thornton not the society. And the version of the counterfactual test set out in the previous paragraph supports that decision.

Two final points should be made. The first is that it was submitted by counsel for Grant Thornton, Simon Salzedo QC, that, if its primary submission on the SAAMCO scope of duty was rejected, the net loss of £26.7m was in any event irrecoverable because the breach of duty did not legally cause that loss. For essentially the same reasons as those given by Teare J, at paras 140-149 of his judgment, I consider that the breach of duty was an effective legal cause of the loss and that the chain of causation was not broken by an intervening event or action. Secondly, what I have here set out is subject to there being a reduction for the contributory negligence of the society both in entering into 50-year swaps, greatly exceeding the likely duration of the mortgages, and in considering that hedge accounting was available when it was not. Teare J thought that the relevant contributory negligence should be 50% (para 255) and there has been no appeal against that figure.

Since writing this judgment, I have had the benefit of reading the joint judgment of Lord Hodge and Lord Sales. It can be seen that our reasoning on SAAMCO is closely aligned not least in relation to avoiding a causation explanation of the SAAMCO principle, the flexible role of the counterfactual test and the importance of the purpose of the duty in determining the scope of the duty of care (although, like Lord Leggatt at para 168 and Lord Hodge and Lord Sales at para 38, I regard the misrepresentation explained in para 206 above as being critical on these facts). While I have stressed that the decision on the scope of the duty of care is underpinned by the policy of achieving a fair and reasonable allocation of the risk of the loss as between the parties, I do not see that as representing a significant difference between us. Where we fundamentally differ is that, with respect, I do not consider it necessary or helpful in this case or in Khan v Meadows - as I explain more fully in my judgment in that case at paras 78-81 - to depart from a more conventional approach to the tort of negligence which begins with the duty of care, treats the SAAMCO principle as being concerned with whether factually caused loss is within the scope of the duty of care, avoids the novel terminology of the “duty nexus”, and sees contributory negligence as one of several possible defences. I would emphasise that there is no dispute in this case that a duty of care was owed by the defendant to the claimant as regards pure economic loss, that there has been a breach of that duty of care, and that the loss in question has been factually caused by the breach. The question that is being focused on is whether the loss in question was within, or outside, the scope of the duty of care (ie within, or outside, the SAAMCO principle).

It is for these reasons that, in my view, the appeal should be allowed in this case. The overall net loss of (approximately) £26.7m is recoverable, subject to a 50% reduction for contributory negligence by the society.