A PERSISTENT question in the tort of negligence is whether defendants who have breached their duty of care should be held responsible for every aspect of the difference between the claimant's position before and after the breach. The law supposedly offers restitutio in integrum – restoring claimants so far as achievable through money to where they were before the tort – but defendants often succeed in evading responsibility for part or all of the loss by using a variety of arguments or metaphors: “the chain of causation” was “broken”, the loss was “outside the scope of the duty” or the “type of loss” was “not reasonably foreseeable”. It is often difficult to disentangle these points from another sort of point, that part of the loss should be discounted because it would have happened anyway, so that awarding damages for it would make the claimant better off than before the breach. The nomenclature used to describe these points is itself contested – are they about “causation”, legal or factual, or “remoteness” or the “scope of the duty”?

The Supreme Court returned to these issues in Manchester Building Society v Grant Thornton [2021] UKSC 20, a companion case to Meadows v Khan [2021] UKSC 21 (noted McGrath [2021] CLJ 440). Regulators require banks and building societies to hold minimum levels of reserves. The minimum level increases the more volatile those reserves are. Manchester Building Society (“MBS”) made equity release loans that created a prospect of extra profit if market interest rates fell but a risk of loss if market interest rates rose. MBS reduced its risk through swap transactions, in which MBS made offsetting gains if rates rose, accepting corresponding reductions in profit if rates fell. Originally, the accounting treatment of these transactions allowed MBS to ignore the swaps’ volatility. But after MBS issued marketable securities, different accounting rules applied, under which the swaps’ volatility did count. MBS's accountants, Grant Thornton (“GT”) had a solution. It advised that the new rules allowed MBS to match each loan to a swap so that MBS could offset the swaps’ volatility using contrary movements in the loans’ asset value, a process called “hedge accounting”. But GT was mistaken. “Hedge accounting” was ineligible, largely because the loans and swaps were mismatched – the loans lasted for the borrowers’ lifetimes, the swaps for fixed terms of 50 years. Consequently, MBS's reserves fell below the minimum required, a situation it eventually resolved by ending the swaps prematurely, at a cost of £32,682,610 (including transaction costs of £285,460). MBS sued GT in negligence.

At first instance ([2018] EWHC 963 (Comm)), Teare J. found GT liable but awarded damages only for MBS's transaction costs, reduced by 25 per cent for contributory negligence (a figure he would have increased to 50 per cent on the losses for which he denied recovery). The Court of Appeal (Hamblen L.J., Males L.J. and Dame Elizabeth Gloster [2019] EWCA Civ 40) rejected MBS's appeal but disagreed with Teare J.'s reasons, particularly his reliance on “assumption of responsibility”, deploying instead Lord Hoffmann's distinction between “advice” and “information” (South Australia Asset Management Corpn v York Montague Ltd. [1997] A.C. 191 (“SAAMCO”)), despite Lord Sumption's doubts about the “descriptive adequacy” of those terms in Hughes-Holland v BPE Solicitors [2017] UKSC 21. The Supreme Court reversed the Court of Appeal, awarding MBS an additional £13.4m (50 per cent of the losses from breaking the swaps).

A majority of the Supreme Court (Lord Hodge and Lord Sales, with whom Lords Reed, Black and Kitchin agreed) arrived at the result by a rather startling route. They proposed a new analysis of the tort of negligence. Instead of asking whether the defendant's conduct was covered by the duty to take reasonable care, whether the defendant had breached that duty and whether the breach caused the loss, both factually and legally, the majority posed six questions, normally but not inevitably to be asked in the following order [at 6]:

The majority answers Question (1) in the affirmative (at [7]) – pace all previous doubts about the actionability of pure economic loss. It also assumes that the answer to Questions (3) and (4) is “yes” and that the answer to Question (6) is “no” except for contributory negligence. And so, the case turns on the “scope of the duty” and “duty nexus”.

The “scope” question, although presented as answerable simply by knowing “the law”, quickly reveals itself as a normative exercise: the issue is whether the law should allocate the risk to the defendant or the claimant. That is a useful clarification, although one might comment that the language of “scope” is not illuminating: it would be much clearer simply to refer to “risk allocation”. But the question is how to answer it? The majority's choice of examples is perhaps unfortunate, beginning with the confusing Spartan Steel v Martin [1973] Q.B. 27, described (at [8]) as a case in which the duty was imposed “in order to” protect the claimant from suffering damage to its property but not damage consisting of “distinct loss of business”. Given that the majority decides that pure economic loss counts unproblematically as actionable, precisely why the former but not the latter attracts the duty is now unclear.

Nevertheless, the majority offers more clarity in one type of case (at [13]): “[T]he 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).”

The difficulty, however, is what happens when the giver and receiver of advice differ about its purpose. In MBS, GT could claim that it gave the advice purely to help MBS's members understand MBS's true financial position, whereas MBS would say that the purpose was to reduce its need to carry regulatory capital reserves. The majority's reference to “an objective basis” suggests an answer steeped in contractual objectivity, in which statements mean what a reasonable person in the position of the receiver, not the utterer, of words would think. That works in MBS, and justifies the result, but it produces the opposite conclusion in Meadows, in which the same judges give the doctor's meaning priority over the patient's meaning. And the majority offers no explanation of how its approach explains Caparo v Dickman [1990] 2 A.C. 605, whose result it endorses (at [8]; see also Meadows, at [34], [38]), since potential investors in a company under a duty to publish its accounts on the Internet (see Companies Act 2006s 430) might reasonably think that the purpose of the auditors’ duty extends to them.

The majority deprecates any reference to “advice” and “information”, going beyond Lord Sumption in BPE, although if understood as shorthand for investigating the purpose of the defendant's statements, those words were arguably more helpful than an inconsistent use of the term “objective”.

The “nexus” question is perhaps the most puzzling in the new set of six. It arises out of difficulties the courts created in SAAMCO about cases in which the risk allocation issue interacts with a plausible argument that part of the claimant's loss would have happened anyway. The SAAMCO market fall is the clearest example. A principled way forward would be to ask two separate questions: (1) By how much is the claimant worse off from the breach alone and (2) should the law allocate the risk of that loss to the defendant or the claimant? So in SAAMCO, the first step would be to compare the value to the claimant of the loan assuming that the defendant's valuation of the building offered as security was correct with the value at the same time of the same loan using the real, lower value of the building. In both scenarios, the claimant would face the risk of a future market fall in the value of the security, but the two loans would already have different values because of the different starting values for the building. The question is how to estimate that difference. The most straightforward method is to ask experts to value the two versions of the loan assuming that both were on offer as securitised assets at the time they were made. Unfortunately, courts have employed methods, including the SAAMCO process of starting with the whole loss including the market fall and deducting the result of asking the “SAAMCO counterfactual” (“what would the loss have been in a scenario in which the valuation was correct?”), that conflate the two questions and inconsistently compare valuations at the start of the transaction with valuations at its end (see Janet O'Sullivan, “Negligent Professional Advice and Market Movements” [1997] C.L.J. 19). MBS does little to clear up this mess. It appears merely to say that one asks the “scope” question, then the somewhat mysterious “nexus” question and then, as a “cross-check” one carries out the “SAAMCO counterfactual” calculation to ensure that the “nexus” question has not led the court astray.

The central difficulty seems to be failing to ask what effect X has on Y controlling for Z. Courts ask a historical question, “would Y have happened if X had not happened?”, and a normative question, “to what extent does the fact that Z has happened exonerate the defendant?”. But neither of these establishes what the claimant lost as a result of the breach alone.

Lords Leggatt and Burrows concur in the result but dissent from the majority's reasoning. Both regard the six new questions as unnecessary. Lord Leggatt insists that conventional ideas of legal causation suffice. Lord Burrows largely agrees but adds that we should recognise that the existing law, especially the SAAMCO version of “scope of the duty”, rests on normative decisions about risk allocation, not on facts about causation.

Whether the majority's six questions survive remains unclear. Some might see this case as a new start – as a new Donoghue – and certainly its division of the “duty of care” into two distinct issues, actionable loss and risk allocation, has attractions. But doubts must remain about whether the “nexus” question makes sense and about whether it clarifies or further confuses the relationship between what the claimant has lost and how the claimant lost it.