Grant Thornton UK LLP (GT) appeals against an order of Bryan J awarding damages of just over £22.36 million to AssetCo plc (AssetCo) for the negligent audit of its accounts for the years ended 31 March 2009 and 2010. Bryan J gave his reasons in a judgment dated 31 January 2019: [2019] EWHC 150 (Comm), [2019] Bus LR 2291.

GT admitted that it had in important respects carried out the audits in breach of its duty of care and that it had failed in its duty to identify management fraud, particularly dishonest representations and evidence provided to it by senior management in the course of the audits. GT was also the subject of disciplinary proceedings as regards these breaches of duty, in which it accepted that its conduct had fallen significantly short of the standards reasonably to be expected of it.

Although GT admitted breaches of duty, a large number of elements of AssetCo’s claim were in dispute before Bryan J. The trial lasted 20 days, involving extensive evidence from factual and expert witnesses and consideration of a large volume of documents and of 877 pages of written submissions as well of course as oral submissions. The bulk of the evidence went to the counterfactual situations which AssetCo alleged would have occurred in 2009 and 2010 if GT had conducted competent audits. The judge delivered a 493-page judgment with over 1270 paragraphs. He found that the counterfactual situations would have occurred, that the damages sought were within the scope of GT’s duty and were not too remote, and that

Many of the issues before the judge do not arise on this appeal. Factual causation, on a “but for” basis, is not in issue before us. There is no challenge to the judge’s conclusions on the remoteness of loss or mitigation. Some, but by no means all, of the case on the counterfactual situation is not in issue. AssetCo does not challenge the decision on contributory fault and GT, which argued for a greater reduction in damages, was refused permission to appeal against it.

GT’s appeal proceeds on three grounds. First, it submits that the judge should have held that AssetCo had failed to establish that the losses for which it claimed damages were within the scope of GT’s duty of care and that its breaches of duty were the legal cause of those losses. If it succeeds on this ground, the whole award of damages would be set aside. Second, he erred in his application of the principles for awarding damages for loss of a chance in finding that the counterfactual situation was established on a 100% basis, and also erred in his assessment of the chances of four specific matters. Third, the judge failed to give credit for benefits received by AssetCo. If successful, this would eliminate or reduce the losses for which damages were recoverable.

Although AssetCo’s claim relates to the audit of the accounts for the years ended 31 March 2009 (FY2009) and 31 March 2010 (FY2010) (the 2009 accounts and the 2010 accounts respectively), and the judge made findings in respect of both audits, the appeal has been concerned only with the audit of the 2009 accounts. The parties accept that, in the light of uncontested findings made by the judge, the 2010 audit and

AssetCo was the holding company of a group carrying on businesses related to fire and rescue services. Its shares were listed on the Alternative Investment Market (AIM) of the London Stock Exchange.

The group’s businesses in FY2009 were organised in three main divisions, of which the largest and most important was Integrated Support Services. This principally comprised two contracts with public authorities. The first was a 20-year Private Finance Initiative contract between AssetCo London Limited (AssetCo London) and the London Fire and Emergency Planning Authority (LFEPA) dated 16 November 2000 (the London Contract). AssetCo London and other subsidiaries (the London Group) provided, serviced and maintained the London Fire Brigade’s fleet of fire engines and ancillary equipment. Second, under a 20-year Public Private Partnership contract dated 19 April 2006 (the Lincoln Contract) with Lincolnshire County Council, AssetCo Lincoln Limited (AssetCo Lincoln) and other subsidiaries (the Lincoln Group) provided similar services for the Lincoln Fire Brigade. In FY2009, the support services division contributed over 80 per cent of the group’s reported operating profit of £17.43 million and of its reported pre-tax profit of nearly £11.3 million. The London Contract represented the major contributor to these results, to the extent of about 85 per cent of the division’s reported profits.

The great majority of the losses claimed by AssetCo in these proceedings relate to the financial support provided, directly or indirectly, by AssetCo to AssetCo London and AssetCo Lincoln and other subsidiaries connected with the London and Lincoln Contracts. Although both Contracts appeared from the 2009 and 2010 accounts to be profitable, they were in fact heavily loss-making. The Lincoln Contract was terminated in April 2012 and the sub-group dealing with the London Contract was sold for £1 in August 2012.

The other divisions were specialist equipment and vehicle assembly. The specialist equipment division included two genuinely profitable companies, AS Fire and Rescue Equipment Limited (AS Fire) and Todd Research Limited (Todd), which manufactured and distributed safety, cutting and security equipment. These companies were sold in December 2010. The vehicle assembly division reported much reduced profits in FY2009, following one of its operating companies being placed in administration in December 2008. Another operating company in the division went into administration in January 2010.

A separate and successful business, for which preliminary steps were taken in FY2009, involved the provision of fire and rescue services in the United Arab Emirates (UAE). During 2008, AssetCo sought business opportunities in Abu Dhabi. AssetCo (Abu Dhabi) Limited (AADL) was formed as a subsidiary of AssetCo in

The 2009 accounts presented a picture of a successful, expanding and increasingly profitable group. They headlined that, over the year, pre-tax profits had increased by 25 per cent to £11.3 million and earnings per share had increased by 33 per cent to 14 pence. The consolidated balance sheet showed total net assets of £51.56 million. The gross assets included goodwill of just over £57 million, of which £44.2 million was attributed to the support services division and therefore largely to the London and

The picture presented by these accounts was entirely false. In fact, as was common ground, the group was insolvent. It had experienced significant cash flow difficulties in FY2009. Its operations were, with minor exceptions, cash negative and unsustainable without external funding. The management’s statement that the group had adequate resources to continue as a going concern for at least 12 months from the time of signing the 2009 accounts in June 2009 was untrue. The carrying values of goodwill and other fixed assets in the consolidated accounts and of the investment in subsidiaries in AssetCo’s balance sheet were unsupportable. An impairment of at least £66.4 million was required in the goodwill and other assets of the support services division alone. Overall, the group’s assets were overstated by some £120 million.

It is common ground that John Shannon (the Chief Executive Officer) and Frank Flynn (the Chief Financial Officer), acted dishonestly in the preparation of the 2009 (and 2010) accounts and in the representations made by them to GT. It is also common ground that Mr Shannon and Mr Flynn acted dishonestly in relation to the conduct of the group’s business, but the extent of the common ground as regards this dishonest conduct requires examination on this appeal.

It is the duty of a company to prepare its accounts and it is the duty of the auditors to audit those accounts. Mr Shannon and Mr Flynn knew that the 2009 (and 2010) accounts prepared by them were false. They duped GT into giving an unqualified audit report by lying and by forging documents. As the judge recorded at [6], it was common ground that during the audit process “management made dishonest statements to GT, provided GT with fabricated and massaged evidence and dishonestly misstated reported profits, and provided GT with flawed and dishonest forecasts and cash flow projections”.

By way of example, Mr Shannon and Mr Flynn told GT that the “unitary payments” due under the London Contract had increased by nearly £47,000 per month (£564,000 pa) from April 2009 and produced documents to establish it. The statements were dishonestly made, and the documents were forged. It was only on the basis of these alleged payments that the London Contract appeared to be profitable. Moreover, on the basis of these supposedly increased payments, capital expenditure under the London Contract was treated as the sale of a finance lease, enabling alleged future revenue of nearly £7.59 million to be recognised in the 2009 accounts (and at least £13.1 million in the 2010 accounts). Another example was that the 2009 accounts reported bid costs for contracts of over £4.5 million within intangible assets. Such costs could be recognised in this way only if it was probable that the contracts would be obtained. Mr Shannon and Mr Flynn falsely and dishonestly reported that they would probably be obtained. Further, the false forecasts and cash flow statements meant that goodwill and other assets were not impaired as they should have been, and they enabled the accounts to be prepared on a going concern basis.

GT accepted that it should not have given an unqualified audit report on the 2009 and 2010 accounts. It also accepted that if it had applied appropriate professional scepticism and competence, it would have uncovered many, if not all, instances of deceit of the GT audit team by the senior management in the course of the audits of the 2009 and 2010 accounts. It would have concluded that the results and assets of the group and AssetCo were unsupportable. It accepts that, in the 2009 accounts, the value for goodwill should have been reduced by between £43.8 million and £61.6 million and that AssetCo’s investment in subsidiaries should have been reduced by between £20.5 million and £33.5 million.

GT specifically admitted the allegation in the particulars of claim that, if it had acted with proper professional skill and care, the “business of the company would have been revealed as ostensibly sustainable only on the basis of dishonest representations and/or unreasonable positions made and taken by management”. As this admission makes clear, the business was not in truth sustainable.

The true state of affairs became apparent in 2011. New management was appointed in March 2011. Short-term financial support was provided by NAV, which had led the group of investors in AADL in January 2009, while the prospects for the group were assessed. A scheme of arrangement with AssetCo’s creditors under Part 26 of the Companies Act 2006, coupled with a restructuring of its capital, in September 2011 secured the solvency and survival of AssetCo, averting the only alternative of an insolvent liquidation. AssetCo retained the SOC Contract and the group has successfully focused on its business in the Gulf.

AssetCo’s case at trial was that, if GT had performed its duties as auditor competently in 2009, the same sequence of events would then have occurred as in fact occurred in 2011. I will refer to this, as did the judge, as the Counterfactual. GT contended that the Counterfactual would not have occurred and that AssetCo failed to prove any loss. The judge held that the Counterfactual would have occurred and that the chances of it not doing so were either non-existent or so small that no discount should be made to the award of damages.

Instead, the group remained under the control of the existing management, who in particular continued to cause substantial and irrecoverable amounts to be provided to the loss-making subsidiaries.

AssetCo claimed damages for the following losses. First, a total of £23,348,675 was claimed in respect of sums provided to its loss-making subsidiaries between 31 March 2009 and 29 September 2011. These sums were provided by AssetCo from a group cash pool facility with the group’s bankers and represented loans to the subsidiaries. These loans were not repaid, and the judge found at [1239] that they were at all material times irrecoverable. The loans were in part funded out of £7.5 million, plus interest of £235,089, subscribed for preference shares in AADL under the PSA and out of an advance payment of just over £3.6 million made under the SOC contract. These payments were made in breach of the PSA agreement and formed an alternative claim for damages. It was common ground that this claim overlapped with the claim in respect of the loans to loss-making subsidiaries. Second, profits of £1,435,817 made by AS Fire and Todd, which would have been available to AssetCo by way of dividend or otherwise, were used to support the loss-making subsidiaries. This second category does not essentially differ from the first, being simply another means by which AssetCo provided support to trading subsidiaries to cover their losses. Third, AssetCo itself incurred expenditure totalling just over £3.53 million which it would not have spent in the Counterfactual. This comprised just under £820,000 spent on management fees to AC Management Services, a company connected to NAV, between 24 July and 31 December 2009, and just over £2.7 million spent between Mr Davies’ appointment on 23 March 2011 and 7 July 2011. Fourth, £1.5 million was claimed in respect of a payment made by AssetCo in December 2009 under a fraudulent related party transaction with a company connected with Mr Shannon, Jaras Property Development Limited (Jaras). Fifth, just under £1.65 million was claimed in respect of dividends paid by AssetCo in September 2009 and November 2010.

The judge awarded damages in respect of all these claims, except for dividends and, because it was an alternative claim, for the PSA monies and SOC Contract advance payment. He rejected the claim in respect of dividends on the grounds that the board’s decisions to pay the dividends were in the circumstances reckless and constituted in each case an actus novus interveniens. AssetCo does not challenge this decision. As earlier noted, the judge reduced the total amount of damages from £29.8 million to £23.36 million on grounds of contributory fault. Neither party now challenges the reduction, either in principle or in amount.

Ground 1 of GT’s grounds of appeal challenges the judge’s conclusion that the losses claimed by AssetCo fell within the scope of GT’s duty and that its admitted breaches of duty were the legal or effective cause of those losses.

Ground 1 is sub-divided into four paragraphs. The first is that the judge erred in

These challenges to the judge’s reasoning, particularly ground 1(1), make it necessary to look at the way AssetCo formulated its case and at the areas of common ground before the judge, before going in detail to the judgment itself. As ground 1 was developed in argument, a central feature of the submissions of Mr Salzedo QC for GT was that the judge had misunderstood the common ground as to the conduct of the management of the AssetCo group both before and after the 2009 audit. This in turn led the judge to find in error the necessary connection between GT’s admitted failings in the performance of its audit duties and the losses claimed by AssetCo.

The particulars of claim were amended a number of times, but I will refer to the version before the judge simply as the particulars of claim. GT’s alleged breaches were pleaded at paragraphs 11-51. I set out below a summary of the alleged, and largely admitted, breaches as regards the FY2009 accounts. All the breaches concern (i) the misstatement of the true financial state of AssetCo and the group and of the losses being made by the group, by including non-existent profits, overstating the value of assets and treating the group as a going concern, and (ii) the failure of GT to detect that management was dishonestly misrepresenting the position to the audit team. AssetCo summarised the effect of these breaches in paragraph 52:

If GT had properly performed its duties, “[t]he business of the company would have been revealed as ostensibly sustainable only on the basis of dishonest representations and/or the unreasonable positions made and taken by management” (para. 58(1)). In that event the Counterfactual would have occurred in 2009, AssetCo would not have suffered the losses which ensued between June 2009 and September 2011, and AssetCo would have survived, having successfully restructured its capital and compromised the claims of creditors under a scheme of arrangement on the same terms as in 2011.

In paragraph 61 of the particulars of claim, AssetCo alleged that, in consequence of

It is unnecessary to look at GT’s defence, because at the judge’s direction the parties prepared a document entitled Revised List of Issues, which contained a lengthy section entitled Main Areas of Common Ground. A number of points in this section should be noted.

First, it was common ground that management had deceived the GT audit team.

Paragraph 5 is important because it sets out the common ground as regards GT’s breaches of duty. The existence and extent of GT’s negligence was not in issue before the judge. Paragraph 5 stated:

As regards both the 2009 and the 2010 Audits, GT did not request and/or obtain and/or treat evidence supplied by AssetCo plc’s management to support proposed finance lease accounting treatment with appropriate professional scepticism.

As regards both the 2009 and 2010 Audits and the treatment of the impairment of assets:

GT breached its duty to apply appropriate professional scepticism and care to the treatment of the impairment of assets by erroneously concluding that no impairment existed;

GT failed to obtain sufficient audit evidence to support the treatment of impairment of assets;

Had GT not breached its duties in this respect, it would have concluded that AssetCo plc’s goodwill should have been impaired by between £43.8m - £61.6m in FY09 and £36.5m - £56.2m in FY10, and that its investment in subsidiaries should have been impaired by between £20.5m - £33.5m in FY09 and £11.9m - £25.4m in FY10 (see joint statement of expert auditors, para 2.12).

As regards both the 2009 and 2010 Audits, GT breached its duty by failing to require that AssetCo plc’s financial statements refer to the fact that certain funds held by the group were restricted pursuant to the terms of the Preference Share Agreement.

As regards both the 2009 and 2010 Audits, although GT disputes aspects of AssetCo plc’s case in respect of the going concern basis and GTs audit review of it, it is common ground that GT failed to obtain sufficient audit evidence in respect of going concern.

As regards both the 2009 and 2010 Audits, although GT disputes aspects of AsserCo plc’s case in respect of the treatment of intangible assets, it is common ground that GT breached its duties to AssetCo plc in respect of intangible assets and that had it not done so it would have concluded that AssetCo plc’s management was falsely and dishonestly and/or unreasonably seeking to recognise costs as intangible assets, contrary to the true position, so as to inflate the company’s asset position and profits.

As regards the Jaras and Graphic Traffic transactions, GT failed to apply appropriate professional scepticism and failed to obtain sufficient appropriate audit evidence. Had GT acted competently, it would have concluded that these transactions were fraudulent transactions designed to benefit Mr Shannon personally.” [This applied only to the 2010 accounts.]

I turn to the judgment. At [6]-[8], the judge said:

“It is common ground that in those years the senior management team at AssetCo behaved in a way that was fundamentally dishonest. During the audit process management made dishonest statements to GT, provided GT with fabricated and massaged evidence and dishonestly misstated reported profits, and provided GT with flawed and dishonest forecasts and cash flow projections. Outside of the audit process, management were engaged in dishonestly ‘overfunding’ assets (i.e. misleading banks as to the costs of new purchases etc so as to borrow more than was permitted), misappropriating monies, dishonestly under-reporting tax liabilities to HMRC, concluding fraudulent related party transactions and forging and backdating documents.

It is also common ground that at the dates of the 2009 and

GT accepts that it was negligent in a number of respects as the company’s auditor in failing to detect these matters and in giving the company clean bills of health; indeed GT accepts that if it had acted competently (as what has been termed in the proceedings “the competent Auditor”), many if not all of the misrepresentations by AssetCo management would have been discovered. The precise scope of the duties owed by GT and its breaches was not agreed, but the parties agreed in a document produced at my direction, to which I shall return, that to the extent that there was any disagreement on this, that was not material to the dispute.

The points at issue in this case are instead about causation and loss. The bulk of the trial was devoted to the question of whether AssetCo could establish that, had GT acted as the Competent Auditor, events would have turned out as

It can be seen from [6] that the judge identified the “fundamentally dishonest” conduct of the senior management team as falling into two categories. First, as set out in the second sentence, they made dishonest statements and provided fabricated evidence to the audit team. Second, as set out in the third sentence, management behaved dishonestly “[o]utside of the audit process”. In that category, the “overfunding” of assets and the forging and backdating of documents involved borrowings of some £13.3 million in excess of the cost of the assets and the dishonest use of fabricated invoices to mislead lenders. The excess sums were used to sustain the business. The misappropriation of monies is, as I understand it, a reference to using the PSA monies, which should have been retained for AADL, for the general purposes of the group. The fraudulent third party transactions were first concluded in FY2010, being in particular transactions with Jaras and Graphic Traffic Limited for the benefit of Mr Shannon.

The judge summarised GT’s defences at [32], including:

At [61]-[84], the judge set out GT’s submissions as to the content of an auditor’s duty, involving eleven propositions. As a preliminary point, GT submitted that it was important to identify the content of GT’s duty because AssetCo had to show that each type of loss fell within the scope of GT’s duty, i.e. was one that GT owed a duty to prevent.

GT’s seventh proposition was that an auditor’s function was to enable shareholders to have the necessary information to exercise their collective powers, including “to scrutinise the conduct of the company’s affairs and to exercise their collective powers to reward or control or remove those to whom that conduct has been confided” (Caparo Industries plc v Dickman [1990] 2 AC 605 (Caparo) at 630 per Lord Oliver).

GT’s ninth proposition was that “the scope of GT’s duty was to protect the company from the consequences of decisions taken by it on the basis that the accounts were free from material misstatement – but no more”. Its tenth proposition was that “it does not fall within the scope of the auditor’s duty to assume responsibility for general trading losses, or for general business decisions or the fraud or imprudence of management”.

In general terms, as the judge recorded at [86] GT admitted that it had failed to gather sufficient audit evidence and to apply appropriate professional scepticism to that which it did gather. At [87], by way of summary of the common ground between the parties, the judge identified six agreed breaches of duty by GT:

Failures in respect of deliberate overstatement of profit and debtors by way of inappropriate finance lease accounting treatment in relation to the 2009 and 2010 audits, by failing to obtain sufficient evidence and not treating with appropriate scepticism the evidence supplied by management as regards cash inflows and profits, particularly the alleged increase in payments under the London Contract.

Failures in respect of the treatment of impairment of assets, accepting management’s conclusion that there should be no impairment. Applying proper professional standards, it would have concluded that there should be substantial impairments, as detailed above, and that management was dishonestly fabricating the cash flow projections in order to inflate the asset values.

Failures in respect of the reporting of cash balances and breach of the PSA. The accounts should have referred to the restrictions placed on the cash raised under the PSA (the PSA monies).

Failures in respect of going concern. GT failed to obtain sufficient evidence to reach a conclusion on management’s statement that the group had adequate financial resources for at least 12 months from the signing of the accounts or to review important aspects of the cash flow projections used by management. A competent auditor would have identified material uncertainties regarding the group’s ability to continue as a going concern.

GT failed as regards the auditing of management’s capitalisation of bid costs. A competent auditor would have concluded that management was falsely and dishonestly and/or unreasonably seeking to recognise costs as intangible assets, contrary to the true position, so as to inflate assets and profits.

Failures as regards two related party transactions in FY2010, which had the indicia of fraud. A competent auditor would have concluded that these transactions were fraudulent and were designed to benefit Mr Shannon personally.

The failures as regards related party transactions in sub-paragraph (6) are not relevant to the audit of the 2009 accounts and, of the remaining five failures, Mr Salzedo submitted that only three (in sub-paragraphs (1), (2) and (5)) related to dishonesty on the part of management. Sub-paragraph (3) is concerned with the PSA monies. GT itself submitted to the judge that Mr Shannon and Mr Flynn were well aware of the circumstances of this money. It is inconceivable that they did not know that its use was restricted by the PSA Agreement and that any use, save as permitted by the agreement, would be a breach of contract or, as GT submitted, a breach of trust. In FY2009, £3 million of the PSA monies had been used in ways not permitted by the agreement: see the judgment at [131]. I find it hard to see that any such use could be anything other than dishonest on their part. However, Mr Salzedo is right to say, so far as I can see, that the judge did not find that GT should have recognised dishonesty as regards the PSA monies in the course of the 2009 audit: see the judgment at [1041]. Sub-paragraph (4) concerns management’s statement that AssetCo had sufficient resources to continue as a going concern until June 2010. This was premised, at least in part, on the same dishonest representations concerning the unitary payment under the London Contract. Mr Salzedo submitted that sub-paragraph (4) added nothing to sub-paragraph (2), but I reject that. The same dishonesty underpinned two key aspects of the 2009 Accounts: fixed asset values and the going concern basis for preparing the accounts.

In subsequent parts of the judgment, the judge elaborated on some of these breaches of duty. At [103], he said that it was common ground that management made a series of unreasonable and dishonest statements to GT, with the effect that AssetCo was able to continue trading as though it were a going concern. The acceptance of fictitious income on the London Contract and its treatment as a finance lease resulted in the unjustified recognition of significant additional revenue in the FY2009 accounts. At [147], the judge said that the “overall result of GT’s negligence on AssetCo’s financial statements was significant and material. The Group’s assets were overstated by £120 million, profits and cash balances were also greatly overstated, and crucially, AssetCo was said to be able to continue as a going concern, when in fact it was insolvent”. In its written opening for the trial (para. 4), GT had stated that it was not in

In Section I of his judgment, running from [908] to [1045], the judge considered legal causation and scope of duty. Although GT had advanced “legal causation” as a single ground of defence, the judge noted at [908] that in essence it encompassed two conceptually distinct, though linked allegations: first, AssetCo’s losses did not fall within the scope of GT’s duty and, second, its losses were not legally caused by GT’s breach of duty. Although expressed there as a defence, the judge correctly proceeded on the basis that it was for AssetCo to establish that both these requirements were satisfied, as he made clear in the next following paragraph. He summarised the two issues as whether “the type of loss suffered was one from which GT owed a duty to protect AssetCo” (scope of duty) and whether “GT’s breach of duty was a legal or substantial cause of the loss” (legal causation).

At [911] and [918], the judge repeated GT’s ninth and tenth propositions mentioned above as to the scope of its duty as auditor. On this basis, GT submitted that it had no liability for heads of loss comprising the funds provided by AssetCo to its lossmaking subsidiaries, the payments made to such subsidiaries by AS Fire and Todd and the other expenditure incurred by AssetCo. The payment to Jaras in respect of the fraudulent related party transaction occurred during FY2010 and there was no equivalent transaction in FY2009 which GT should have identified.

AssetCo’s submissions were summarised at [912]. First, the losses resulted from GT’s failure to detect “fraud or irregularity likely to result in material losses to the company” (Sasea Finance Ltd v KPMG [2000] BCC 989 at 994). Second, the losses were suffered as a result of AssetCo continuing to trade “in a particular manner” (Temseel Holdings v Beaumonts [2003] PNLR 27 at [52]). AssetCo relied on what the judge said, by reference to GT’s opening skeleton (para. 3), was the common ground that AssetCo was run by Mr Shannon and Mr Flynn in a fundamentally dishonest way and also on the common ground in the Revised List of Issues that AssetCo was “ostensibly sustainable only on the basis of dishonest representations or unreasonable decisions made or taken by its management”. AssetCo submitted that every aspect of its business, including the wasted expenditure claimed as damages, was touched by the fraud of Mr Shannon and Mr Flynn. Third, in those circumstances, all of the losses suffered by it fell within the scope of GT’s duty. Fourth, the losses were legally caused by GT’s negligence and there was no break in the chain of causation, because the losses were the “very thing” that GT was under a duty to protect against (see Barings plc v Coopers & Lybrand (No 7) [2003] EWHC 1319 (Ch), [2003] Lloyd’s Rep IR 566).

As regards the largest part of AssetCo’s claim, the funds provided to the loss-making subsidiaries, the judge set out the parties’ positions as follows:

That is largely agreed by AssetCo, subject to a question about the precise meaning and role of “reliance”, which I will address separately below to the extent that it is relevant (which ultimately it is not, in my view, in relation to the primary debate about scope of duty). AssetCo accepts that it cannot recover losses that were suffered simply because the company remained in existence and carried on trading, but avers that it can recover losses suffered by AssetCo continuing to trade in a particular fashion in reliance on the audit. The dispute is centred, in particular, on how narrowly or broadly the concept of “a particular manner/fashion” should be defined.

AssetCo says that its losses resulted from its business being run in a fundamentally dishonest way, and that, since GT should have detected that AssetCo was being run in that way, AssetCo’s losses were of a type that GT was under a duty to prevent, and indeed were suffered by its trading in a particular (i.e. fraudulent) manner as a result of GT’s negligence. GT’s case is that that is too vague, and that the “particular manner” in which AssetCo conducted its business must be more narrowly defined such that AssetCo can say that each head of loss was caused by a particular fraud which GT was required to, and negligently failed to, identify.”

Between [922] and [955], the judge reviewed in detail the authorities on which the parties principally relied: Bank of Credit and Commerce International v Price Waterhouse [1999] BCC 351 (BCCI) and Equitable Life Assurance Society v Ernst & Young [2003] EWHC 112 (Comm), in the case of GT, and Sasea Finance Ltd v KPMG [2000] BCC 989 and Temseel Holdings v Beaumonts [2003] PNLR 27) and Livent v Deloitte & Touche 2107 SCC 63, a decision of the Supreme Court of Canada, in the case of AssetCo. The judge did not find great assistance from these cases and for the most part they have featured little, if at all, in the submissions to us.

The judge, therefore, considered it necessary to focus on the purpose of an auditor’s duty of care, as explained by the House of Lords in Caparo.

The judge concluded as follows (the quotations italicised by the judge are taken from the speeches of Lord Bridge and Lord Jauncey in Caparo):

The judge said at [963] that GT’s submissions, based on the role of auditors to audit, not to prepare, accounts and their more limited role of investigation into a company as compared with that of the directors “are not in point where it is common ground that GT’s duty was sufficiently broad that GT should, in the proper exercise of that duty, have uncovered many, if not all, of the instances of management deceit carried out by Messrs Shannon and Flynn, consequent upon which AssetCo would have called the directors to book and ensured that the company did not continue to trade (given that the company was “ostensibly sustainable only on the basis of dishonest representations or unreasonable decisions made or taken by management – Revised List of Issues para 6(1)).”

The judge at [964] rejected GT’s submission that this conclusion had the effect that an auditor who fails to identify a particular fraud effectively becomes the insurer of the company for any dishonesty or fraud within the company and trading losses suffered as a result. At [966], he rejected the suggestion that the consequences were draconian or unfair to the auditor and continued:

The judge’s analysis summarised above was in the context of the scope of the auditor’s duty. At [970] he turned to the separate requirement of legal causation, observing that “the law distinguishes between a breach that was the effective cause of

The judge considered reliance at [973]-[982], concluding that, while it was an essential element, AssetCo did not need to show that its board or shareholders subjectively had the audit reports in mind when they made decisions; it was enough that they would have acted differently but for GT’s breach. There is no challenge to that conclusion. As the judge had found the Counterfactual to be established, the requirement for reliance was satisfied.

After examining the law on intervening acts and the “very thing” principle at [983][990], the judge said at [991] that he had to consider whether on the facts GT’s breaches were the legal cause or the mere occasion of the four heads of loss claimed by AssetCo and whether any intervening act had broken the chain of causation.

Dealing first with the support provided to subsidiaries to meet their trading losses, the judge said at [996]:

As regards the Jaras payment, the judge rejected GT’s submission that the 2009 audit provided no more than the mere occasion for this loss. GT accepted that, if there had been a similar fraud in FY2009 which GT should have but did not uncover, its breach would have been the legal cause of the loss in respect of the Jaras payment, but there had been no such fraud in FY2009. The judge held at [1001]:

While that completes the parts of the judgment directly related to the scope of GT’s duty and legal causation, there are further relevant observations by the judge in section K ([1091] - [1190]) where he dealt with contributory fault.

At [1143] and in a number of other paragraphs (for example, [1184] and [1186]), the judge said:

Speaking of the “very thing” principle, the judge said at [1148] that “(to take one example) the fraud in relation to the unitary payment was an important part of the dishonest trading – but it was a fraud that GT has admitted would have been uncovered had it exercised proper scepticism”.

At [1167], the judge said of the Jaras transaction, “this head of loss is intimately tied up with GT’s own breaches of duty”. GT “would have uncovered Mr Shannon’s dishonesty if [it] had performed its duties in the 2009 audit”.

GT challenges the judge’s decision on scope of duty and legal causation as regards all heads of loss, except of course AssetCo’s claim in respect of dividends rejected by the judge. I have earlier summarised its grounds of appeal on these issues.

The first broad ground of challenge relates to the judge’s statement, frequently repeated in the judgment, that the business of AssetCo was being run in a fundamentally dishonest manner. The judge accepted AssetCo’s case that every aspect of the business was touched by the fraud of Mr Shannon and Mr Flynn. By admittedly failing to detect management’s deceit of the audit team in the course of the audit of the FY2009 accounts, the judge held that GT was responsible for all the losses incurred by AssetCo in the course of its business conducted in that fundamentally dishonest manner. GT submits that it was not common ground, nor was it even pleaded by AssetCo, that its business was conducted in a fundamentally dishonest manner.

GT accepts in principle the way the judge put the issue at [920], that AssetCo could not recover losses that were suffered simply because AssetCo carried on trading but could recover losses suffered as a result of “continuing to trade in a particular fashion in reliance of the audit” (the judge’s emphasis). He continued that the dispute was centred, in particular, “on how narrowly or broadly the concept of “a particular manner/fashion” should be defined”. However, GT submits that the judge was wrong to accept AssetCo’s submission that its losses resulted from its business being run in a particular (fundamentally dishonest) way which should have been detected by GT and that GT’s failure to do so deprived AssetCo of the opportunity to correct it, so that such losses were recoverable from GT.

GT submits that the judge’s conclusion on the scope of its duty are based on two key propositions, neither of which was correct on the evidence before him. They were, first, that GT’s relevant breach of duty was a failure to detect that AssetCo was being run in a fundamentally dishonest way and, second, that the losses arose out of the same fundamentally dishonest conduct of the business. However, the judge nowhere identifies the fraud or dishonesty which is said to have caused the group’s losses nor how it operated to cause those losses.

GT further submits that the admitted breaches of duty by GT in the 2009 audit related to a failure to detect dishonest accounting designed to conceal losses and AssetCo’s insolvency by inflating profits and asset values and by making it appear that the group was a going concern. None of them concerned dishonesty or fraud in entering into any transactions. The dishonesty lay in the false accounting, not in the conduct of the business. There was therefore no link found by the judge between the false accounting and GT’s failure to detect it on the one hand and the trading and other losses incurred by AssetCo on the other. Without that link, the losses claimed by AssetCo could not fall within the scope of GT’s duty nor could its breaches of duty be the legal cause of those losses. There were some specific instances of dishonesty in the conduct of the business, such as overfunding, but it was not part of the case as set out in the revised List of Issues that GT should have detected them nor, except for the Jaras payment, were they were part of the claimed losses. GT did not accept, nor was it alleged or proved, that the business was as a general proposition being conducted in a dishonest, or fundamentally dishonest, manner.

There is, in my judgment, substance in these criticisms. There is a lack of clarity and consistency in the way the judge approached the proposition that the group’s business was being run in a fundamentally dishonest way. To be fair to him, I think much of the confusion stems from the way in which the common ground was expressed in the Revised List of Issues. While GT’s admitted breaches of duty were clear and detailed, the same cannot be said of the agreed consequences of the breaches.

The statement of common ground in the judgment at [6] is particularly important. GT does not challenge it, indeed it is taken from GT’s opening written submissions for

In my judgment, Mr Salzedo is correct to submit that [6] is not a statement of common ground that the business was being conducted in a fundamentally dishonest way. Rather, it is a statement that the senior management team’s conduct in the respects particularised in the following two sentences was fundamentally dishonest. The distinction is important. There are very many cases in which it is found that directors or managers have engaged in dishonest conduct in the course of running a business (as, for example, in the ways identified in the third sentence of [6]), without the entire business being run in a fundamentally dishonest way. The latter on the other hand occurs, for example, where the business is a Ponzi scheme or where the business can be continued only by taking on credit which the management knows cannot be repaid. It was not suggested that AssetCo and its group businesses fell into this type of category. There were dishonest transactions, such as the over-funding, but overall the problem with the operation of the main businesses of the London and Lincoln Contracts was that they were heavily loss-making, such that the group was insolvent and could not continue as a going concern.

Moreover, it was not common ground that GT was negligent in failing to detect any of the transactions referred to in the third sentence of [6] (with the exceptions of the use of the PSA monies and the related party transactions in FY2010) nor (with the same exceptions) did AssetCo make any claim for loss arising from those transactions. It may be noted that at [691] the judge said that it was common ground that a competent auditor would have discovered management dishonesty in relation to overfunding, but this appears to be an error: it does not feature as common ground in the Revised List of Issues or in the summary of agreed breaches of duty in the judgment at [87].

I should add that Mr Salzedo took us through the relevant parts of the skeleton arguments and oral submissions before the judge and I am satisfied that GT did not accept that there was generalised fraud in the business such that it could be said to have been run in a fundamentally dishonest way. Counsel also made clear to the judge his submission that a failure to detect specific incidents demonstrating particular types of fraud in the conduct of a company’s business does not expose the auditor to liability for all losses resulting from fraud of all kinds. If, for example, the only negligence of an auditor was a failure to detect a fraudulent transaction with a company connected with a director, the auditor may well be liable for subsequent losses of a similar kind but will not be liable, when it is discovered that the company was operating as a Ponzi scheme, for all the losses resulting from the generally fraudulent conduct of the company’s business. In my judgment, this submission was correct.

Thus far I agree with GT’s challenges to the judge’s reasoning. However, it is by no means a complete statement of the judge’s grounds for holding GT liable for the trading and other losses. The common ground between the parties included what was stated by the judge at [7] and frequently repeated by him in the course of the judgment: “at the dates of the 2009 and 2010 Audits, AssetCo’s business was ostensibly sustainable only on the basis of the dishonest representations or unreasonable decisions made and taken by them”.

The dishonest representations were those made by management to GT in the course of the audits. On the basis of those representations, GT issued unqualified audit reports that the accounts, which showed AssetCo and its group to be profitable, to have very substantial net assets and to be going concerns, gave a true and fair view. The business was in this way “ostensibly sustainable”. In other words, it was made to appear that it was sustainable when in truth it was insolvent. This occurred only because, as GT accepts, it negligently failed to detect that the position was being deliberately misrepresented to it in the respects set out in the judgment at [87]. In the absence of this negligence, GT accepts that the true position would have been discovered and the judge found that the Counterfactual would have occurred.

This was an integral part of the judge’s reasoning and was at least part of his assessment that AssetCo and its group were being run in a fundamentally dishonest way. It is in no sense a misuse of language to say that, where the continued operation of the business was only possible because management deceived GT into reporting, as true and fair, accounts which wholly misrepresented the true financial position, it was being run in a fundamentally dishonest way.

The questions are whether, in those circumstances, the ensuing trading and other losses of AssetCo fell within the scope of GT’s duty and whether its admitted breaches of duty were the legal cause of those losses. These are the questions to which the rest of GT’s Ground 1 is directed.

GT submits that the proper approach to these questions is set out in the Supreme Court’s decision in Hughes-Holland v BPE Solicitors [2017] UKSC 21, [2018] AC 599 (Hughes-Holland) and this court’s decision in Manchester Building Society v Grant Thornton UK LLP [2019] EWCA Civ 40, [2019] 1 WLR 4610 (MBS).

The issue in Hughes-Holland, as identified by Lord Sumption in his judgment (with which the other members of the Court agreed) at [1] was “what damages are recoverable in a case where (i) but for the negligence of a professional adviser his client would not have embarked on some course of action, but (ii) part or all of the loss which he suffered by doing so arose from risks which it was no part of the adviser’s duty to protect his client against”. Translated into the facts of the present case, GT (i) accepts that but for the negligence of GT AssetCo would not have continued in business after its unqualified audit report on the 2009 Accounts (except arguably by way of the Counterfactual) but (ii) the losses which AssetCo thereafter suffered arose principally from the risks of the continuation of a loss-making business, against which GT had no duty to protect AssetCo.

At [20]-[46], Lord Sumption discussed the general principles applicable to scope of duty and legal causation as factors which in law limit liability for professional negligence, by reference to previous authority, in particular the decision of the House of Lords in Banque Bruxelles Lambert SA v Eagle Star Insurance Co Ltd [1997] AC 191 (SAAMCO).

SAAMCO established or reiterated that a claimant must establish that its loss fell within the scope of the duty owed by the defendant to the claimant or, to put it another way, the issue is the extent of the defendant’s duty to protect the claimant against the loss caused on a but for basis by the breach of duty. It also established the importance of determining whether the defendant was providing information or was giving advice generally. Lord Sumption rejected the criticisms which had been made of this distinction. The correct categorisation will determine the extent of the defendant’s liability. If it is an advice case, the defendant has a duty to protect the claimant against the full range of risks associated with entering into the transaction. In an information case, the defendant supplied only a part of the material on which the claimant decided whether to enter into the transaction, and the identification of other relevant considerations and the overall assessment of the merits of the transaction were matters for the claimant: see Lord Sumption at [40]-[41]. In an information case, 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”. The defendant is not liable for consequences which would have occurred even if the information had been correct: see SAAMCO at p.213F and p.214D and Hughes-Holland at [30]. At [45], Lord Sumption said:

The question posed by the SAAMCO principle is designed to direct the analysis to the link between the particular breaches of duty and the particular losses; which of those losses are the result of the wrongness of the information as opposed to the transactions entered into as a causal result of the wrong information.

The effect of these decisions was reviewed and applied by this court in MBS. As part of its business, the claimant building society issued fixed interest lifetime mortgage loans. It hedged its interest rate risk on those mortgage loans (the risk that the variable interest rates that it paid to acquire funds would exceed the fixed rate under the loans) by entering into interest rate swaps. From 2005, the claimant was required by accounting standards to include swaps at fair value in its balance sheet, the fair value being its mark-to-market (MTM) value at the balance sheet date. This exposed the claimant’s financial position, as stated in its balance sheet, to volatility arising from changes in MTM values. This could be mitigated by the use, if permissible, of hedge accounting. The defendant, which was MBS’s auditor, advised in April 2006 that it could apply hedge accounting, which was accordingly adopted in the accounts for each year from 2006 to 2011. As auditor, the defendant gave an unqualified audit report for those accounts, repeating its approval of the use of hedge accounting. From April 2006, the claimant relied on the defendant’s advice to enter into further lifetime mortgages and further long-term swaps. If the defendant had not so advised, it would not have entered into any further swaps and it would have broken the swaps it already held.

In 2013, the claimant was advised that in its circumstances hedge accounting was not, and had not since 2006 been, permissible. The necessary change in accounting treatment turned a profit into a significant loss and substantially reduced its net assets, such that it no longer had sufficient regulatory capital. The claimant closed out the swaps, which were by then heavily out of the money, resulting in substantial losses.

This court, affirming the decision at first instance but on different grounds, held that the defendant was not liable for the losses incurred by the claimant when closing out the swaps (the MTM losses). Hamblen LJ, with whom Males LJ and Dame Elizabeth Gloster agreed, analysed SAAMCO and Hughes-Holland at [46]-[53], finishing by referring to “the SAAMCO cap” or principle and citing the passage in Hughes-Holland at [45] where Lord Sumption referred to it as a tool, which determines the loss resulting from the defendant’s negligence. At [54], Hamblen LJ said that the SAAMCO principle “may generally be addressed by considering the following” six questions. The first step is to consider whether it is an “advice” case or an “information” case, for the reasons which I have earlier set out. If it is an “information” case, the fifth step is that “the negligent adviser/information provider will only be responsible for the financial consequences of the advice and/or information being wrong”. (The reference is made there to “adviser” and “advice” because advice on a particular aspect of a proposed transaction or course of action, rather than advice on the merits of the transaction or course of action overall, makes it an “information” case: see Hughes-Holland at [39]-[41].) The sixth step explains that the fifth step “involves a consideration of what losses would have been suffered if the advice and/or information had been correct. It is only losses which would not have been suffered in such circumstances that are recoverable.”

At [55], Hamblen LJ said that it was clearly a case in which the SAAMCO principle applied and that it was properly analysed as an “information” case. In April 2006, the defendant gave accounting advice on one topic, the applicability of hedge accounting to the treatment of swaps in the claimant’s accounts. By giving an unqualified audit report for each of the years 2006 to 2011, the defendant implicitly repeated the advice it had given in April 2006. It was plain that the defendant’s responsibility was limited to giving accounting advice and never came close to extending to responsibility for the entire lifetime mortgage/swaps business ([70]). The correct question was therefore whether the claimant would have suffered the MTM losses if the accounting advice had been correct. For reasons that are immaterial for present purposes, it was held that the claimant had failed to prove that it would not have suffered those losses.

Relying on these cases, Mr Salzedo submits that (i) the SAAMCO principle applies to claims against auditors, including (as in this case) general audit claims, (ii) such claims are “information” cases, and (iii) it is therefore essential for the claimant to establish that the losses it claims would not have been suffered if the audit reports had been correct. He goes on to submit that in this case AssetCo had not asked the judge to find whether the losses would have been suffered if the audit reports had been correct and the judge had neither asked that question nor provided an answer. AssetCo had therefore failed to establish a basic and necessary element of its claim.

Expanding on these submissions, Mr Salzedo argues that an audit is a statutory requirement involving the provision of an audit report on a company’s accounts. If the audit report is unqualified, it involves numerous implicit statements about the components of the accounts, and the company may rely on one or more of those implicit statements in deciding whether to enter into particular transactions. In addition, and this is central to the present case, auditors are under a duty to use reasonable care and skill to identify and report the existence of fraud and dishonesty. The absence of a report of fraud or dishonesty is an implicit representation that none exists, which is negligently made if there was fraud or dishonesty that the auditor, exercising professional care and skill, should have detected. In the present case, GT admits that it was negligent in this respect.

In applying the authorities to the present case, Mr Salzedo submits that the correct approach is to identify the relevant information that was negligently provided (or not provided) by GT and to ask what part of the claimed losses would have been suffered if the information had been true. The relevant information was that relating to management dishonesty, as summarised in the judgment at [87], all of which as regards the 2009 accounts involved dishonest accounting to create a false picture of profits and greatly inflated net assets.

Mr Salzedo submits that the present is an information case, involving the implicit statement that there was no management fraud or dishonesty in the preparation of the 2009 accounts. It could not sensibly be suggested that GT assumed responsibility for the merits of continued trading by AssetCo.

The relevant question is therefore: if it had been correct that management was not being dishonest in their preparation of the accounts in the ways that were common ground, would the claimed losses still have been suffered? In particular, would AssetCo have continued to provide very substantial sums to support its subsidiaries? The answer, Mr Salzedo submits, is that there is no reason to suppose that the support would not have been provided or that the losses would not have been incurred. There was no finding that the losses incurred in providing support to the subsidiaries was anything to do with management’s dishonesty in the preparation of the accounts. Accordingly, the judge was wrong on scope of duty and there is nothing in the facts found by him to establish that the claimed losses were within the scope of GT’s duty.

By way of preliminary submission, Mr Templeman QC objects that GT had not raised the SAAMCO principle before the judge, although it did cite Hughes-Holland for a different point, and it is too late to raise it now. By contrast, as mentioned above, Mr Salzedo’s submission is that the burden lay on AssetCo to allege and prove this essential part of its case and it failed to do so. In my view, the submissions of both parties misstate the purpose of the SAAMCO principle. It is, as Lord Sumption said in Hughes-Holland, simply a tool for determining the losses which fall within the scope of the defendant’s duty. There may be cases in which a claimant cannot succeed without specifically alleging and separately proving that the loss would not have been suffered if the information provided by the defendant had been correct. But in SAAMCO and in many other cases, and for reasons which will appear I consider the present case to be one of them, the answer is established by the evidence as a whole without any need for it to be separately addressed in the course of the evidence. The fact that the judge was not invited by either party to address it need not, in this case, inhibit this court from doing so. In oral submissions, Mr Salzedo put the point somewhat differently, saying that the underlying principle that recoverable losses must fall within the scope of the defendant’s duty was clearly put to the judge. He was asked to apply the right principle but came to the wrong answer.

Moving to the substance of the point, Mr Templeman submitted that the SAAMCO principle, while a useful and workable test where professional information or advice is provided in the context of a specific transaction or a course of conduct then in contemplation, was not applicable to the very different context of an audit report to a company and its shareholders. In the case of the single transaction, the information or advice will form only part of the material which the recipient will consider before committing to the transaction. It is directed to the recipient taking a decision. In those circumstances, the provider of the information is liable only for the consequences of that information being wrong, determined by asking whether the loss resulting from the transaction would have been suffered even if the information had been correct. Mr Templeman drew attention to the numerous occasions in the judgment of Lord Sumption in Hughes-Holland when he referred to the principle as applying in the case of a specific transaction or course of action. The same is true of MBS. Although the relevant advice was given by MBS’s then auditors, it was directed specifically to whether MBS should continue to issue fixed rate mortgage loans hedged by interest swaps. It was given in the context of a specific proposed course of action and the subsequent audit reports were relevant only because they implicitly approved the advice previously given. Indeed, Hamblen LJ said at [50] that the SAAMCO principle provides a filter to eliminate certain losses from the scope of the defendant’s negligence “in cases where foreseeable losses are suffered as a result of entering into a transaction in reliance on negligent advice and/or information” (emphasis added).

Mr Templeman submits that, by contrast, in the case of an audit report, there is no specific transaction to which it is directed. It makes no sense to ask whether the professional person was only providing information or was assuming responsibility for the decision of the recipient to enter into the transaction. In such circumstances, the court applies the usual rules of factual causation, scope of duty, legal causation and remoteness, without any need to add the further complication of the SAAMCO principle.

The purpose of an audit report, Mr Templeman submitted, is entirely different to the provision of information or advice to be used by the recipient in deciding whether to enter into a specific transaction or embark on a specific course of action. As explained in Caparo, the purpose of an audit report is to enable the company, either the company in general meeting or the non-executive directors, to hold the management of the company to account and to ensure that errors in management are corrected. If the audit report fails to disclose, for example, management fraud and dishonesty, the result is that the company is deprived of the opportunity to take the action that otherwise it would have taken, to rectify the situation.

Mr Templeman relies on the judgment of Evans-Lombe J in Barings plc v Coopers & Lybrand [2003] EWHC 1319 (Ch), [2003] Lloyd’s Rep IR 566. A subsidiary in the Barings group (BFS) incurred large losses as a result of unauthorised securities trading in Singapore which were concealed by a dishonest employee. BFS received financial support from other group companies which it was unable to repay. EvansLombe J held that BFS succeeded in its claim against its auditors for negligence in the conduct of the audits of BFS’s accounts for two years and awarded damages for the trading losses incurred in an 18-month period.

Evans-Lombe J said:

Lord Hoffmann’s test, to enquire whether the losses funded by the Dollar Funding “would have occurred even if the information which [D&T] gave had been correct” does not assist. If the 1993 audit certificate had been correct, it would mean that there had been no unauthorised trading up to 31 December 1993. In those circumstances it must be most unlikely that Leeson would have started unauthorised trading in 1994. It would be unwarranted speculation to assume that Leeson would have done so.

Nonetheless a reasonable “common sense” case can be made for saying that D&T should not be liable for loss occurring after the commencement of the improvident Dollar Funding, of which D&T could have known nothing and from which they could be under no duty to protect BFS.”

Mr Templeman also relies on the decision of Fancourt J in BTI 2014 LLC v PricewaterhouseCoopers LLP [2019] EWHC 3034 (Ch) in which he refused the defendant auditor’s application to strike out a claim for damages in respect of dividends paid on the basis of accounts said to have been negligently audited. It was alleged that competently audited accounts would have shown that the company did not have distributable reserves. The auditors argued that, applying the SAAMCO principle, the loss represented by the dividends was outside the scope of their duty. If the audit report had been accurate, the company would have had the necessary distributable reserves and it would not have suffered the alleged loss. Fancourt J considered that to be intuitively wrong, as it would always absolve auditors from liability for dividends unlawfully paid as a result of their negligent audit. It was also contrary to authority going back to the 19^th Century.

It is right, as Mr Templeman has done, to draw attention to the significant differences between the position of auditors issuing an unqualified audit certificate and the position of a professional who is just providing information or advice on a particular aspect of a proposed transaction. An auditor, unlike a professional in the latter case, is performing a statutory duty which relates to all the matters which the directors are required to report in their company’s accounts. The purpose is not to assist the recipient to decide whether to enter into a particular transaction or pursue a particular course of action.

The purpose is that expressed in the speeches in Caparo. Lord Bridge said ([1990] 2 AC 605 at 626):

At p. 630, Lord Oliver said: “It is the auditors’ function to ensure, so far as possible, that the financial information as to the company’s affairs prepared by the directors accurately reflects the company’s position in order, first, to protect the company itself from the consequences of undetected errors or, possibly, wrongdoing (by, for instance, declaring dividends out of capital) and, secondly, to provide shareholders with reliable information for the purpose of enabling them to scrutinise the conduct of the company’s affairs and to exercise their collective powers to reward or control or remove those to whom that conduct has been confided.”

Mr Salzedo argues that Caparo was concerned with the persons to whom the auditors owed their duties and the kind of transaction to which those duties related or, as Lord Bridge put it, “the extent of the interest which the auditor has a duty to protect”. While he is right about that, it does not diminish the relevance and importance of the statements about the purpose of an audit in determining the scope of the auditor’s duty to the company. As Mr Salzedo accepts, when shareholders exercise their powers to remove the directors or to influence the conduct of the company’s affairs, they do so not in their personal capacity but as an organ of the company. As he also accepts, it need not be the shareholders that take action; it may, for example, be the nonexecutive directors.

Nor do I think that Mr Salzedo can rely on MBS to support his case. The claim there was focused on a particular piece of advice given in the context of whether to continue a particular line of business. The advice could have been given by accountants other than the auditors; a company with a small firm as auditors might seek and rely on advice from a large or specialist firm on a difficult question of accounting treatment and the application of accounting standards to it. It was not concerned with the scope of an auditor’s duty when giving an audit report.

Nonetheless, I do not see any substantial reason why the SAAMCO principle cannot or should not, in most circumstances, be applied to determine whether particular losses come within the scope of the auditor’s duty when signing an unqualified audit certificate. Its purpose, as Phillips LJ said in the course of argument, is to distinguish the negligent audit that is merely the occasion for the loss from the negligent audit that gives rise to a liability to make good the loss. It is capable of being effectively applied to most types of loss that may be claimed in respect of a negligent audit. As was also observed in argument, Lord Hoffmann made extensive reference to Caparo and the position of auditors in his speech in SAAMCO, without any suggestion that the principle he developed in that case did not apply to claims against auditors.

There may be exceptions to the application of the SAAMCO principle. The case of dividends which could not lawfully have been paid if the accounts had been competently audited is an obvious candidate. If the question is asked whether the loss would have been suffered if the accounts and the audit certificate had been correct, the answer has to be in the negative but that produces a result that is not only intuitively wrong but contrary to a long-established line of authority that holds auditors liable in such circumstances, as Fancourt J pointed out in BTI 2014 LLC v PricewaterhouseCoopers LLP. However, accepting that there are or may be situations in which the SAAMCO principle may be inapplicable, it does not follow that it is not generally applicable to the liability of auditors. It is not a rigid rule of law but, as Lord Sumption in Hughes-Holland at [45], “simply a tool” for determining the loss flowing from the negligently wrong information as opposed to the loss flowing from entering into the transaction at all. If, in a particular class of case it is incapable of achieving that determination, it is not a tool which the court will use.

As regards the observations of Evans-Lombe J in Barings, I agree with Mr Salzedo that the SAAMCO principle was capable in that case of producing the correct answer. Moreover, it did in fact produce the right answer. As Evans-Lombe J said at [818], the SAAMCO principle produced the answer that, if BFS’s accounts and the audit certificates had been correct, there was no basis for supposing that BFS would have suffered the losses resulting from the unauthorised trading. As he said, it would be unwarranted speculation to assume that unauthorised trading would start in the following year. In saying that the SAAMCO principle did not assist, Evans-Lombe J was referring to its inability to assist the auditors in their defence. The issue addressed by the judge was whether the auditors were liable for losses incurred after the start of and as a result of “Dollar Funding”, an improvident supply of funding from other Barings companies. While Evans-Lombe J considered that there was a reasonable common sense case for saying the auditors should not be liable for these losses, he concluded that they could not avoid liability save on a basis that might involve reopening their concession that the losses were foreseeable and not too remote: see [825]. I do not read the judgment in Barings as a rejection of the SAAMCO principle in general audit cases and, if it were, I would not consider it to be correct.

Mr Templeman submits that, assuming against himself that the SAAMCO principle applied in the present case, AssetCo’s claim satisfied it. GT’s audit reports stated its opinion that the 2009 accounts gave a true and fair view of the state of the affairs of the group and of AssetCo. GT accepts that these opinions were negligently wrong. GT failed to detect that the senior management were lying to them about the true state of

Mr Templeman submits that, in these circumstances, if one applies the SAAMCO principle and asks whether AssetCo would have suffered the losses claimed if the accounts and audit reports had been true, the answer must be “no”. Applying the same approach as Evans-Lombe J in Barings, it would be unwarranted speculation to assume that the group’s businesses would have become loss-making in FY 2010 or FY 2011. The results shown in the 2009 Accounts were on the false basis that the London and Lincoln Contracts were profitable. If they had been profitable in FY 2009, there is no reason why they should become loss-making in the following two years.

Mr Salzedo submits that there is no sufficient connection with the admitted breaches of duty summarised in the judgment at [87], including those which he accepts involve a failure to detect dishonesty on the part of management in what they were telling the GT audit team, and any losses arising from dishonesty in the conduct of the business in FY2010 and FY2011. For example, there are no claims for losses resulting from overfunding or misleading HMRC, nor indeed do the admitted breaches of duty relate to such misconduct.

As I have earlier indicated, these submissions on behalf of GT leave out of account a major part of the case made against and accepted by GT, namely that at the date of the audit of the 2009 accounts, the business was ostensibly sustainable only on the basis of the dishonest representations or unreasonable decisions made and taken by management. The business was in truth not sustainable and GT was in admitted breach of duty in not detecting the dishonest misrepresentations made to it. GT failed to detect that the senior management was deliberately concealing the true state of the business. It was that failure which caused GT negligently to issue the unqualified audit opinion on the 2009 accounts, rendering the business “ostensibly sustainable”.

This is not just a case of “but for” causation, with the audit certificate providing the occasion for losses subsequently suffered. GT here failed to detect the dishonest concealment of the substantial losses made in FY2009 and the group’s insolvency, which continued in FY2010 and FY2011 resulting in the losses claimed by AssetCo, or most of them. This failure deprived AssetCo (whether acting by its shareholders collectively in general meeting or by its non-executive directors) of the opportunity to call the senior management to account and to ensure that errors in management are corrected. As the speeches in Caparo make clear, this is a principal purpose of an audit and, as Lord Hoffmann made clear in SAAMCO at p.212, the scope of an auditor’s duty is determined by reference to the purposes of the statutory requirement for an audit. Liability for the losses suffered in FY 2010 and FY2011 is “liability for the consequences of the information being inaccurate” (SAAMCO per Lord Hoffmann at p.213).

I have earlier set out [961]-[962] of the judgment, where the judge quotes from Caparo as regards the purpose of an audit and concludes that the trading losses fell within the scope of GT’s duty. I also quoted from [963] where he made, I believe, the

In my judgment, the judge was therefore right to conclude that GT was liable for the losses incurred in supporting its subsidiaries, whether directly or indirectly by payments made by AS Fire and Todd. No separate argument was advanced for treating differently the other operational losses incurred by AssetCo between June 2009 and September 2011.

The claim for loss in respect of the Jaras transaction is, however, in my judgment to be treated differently. It involved Mr Shannon misappropriating company funds for his own personal benefit. There was no similar transaction in FY2009 and no negligence on the part of GT in respect of transactions of this type in its conduct of the 2009 audit. It might be said that directors who are prepared to lie and forge documents to deceive the auditors into accepting the company and the group as solvent and profitable are capable of any fraud in the conduct of the company’s business and it was GT’s negligence that allowed Mr Shannon to perpetrate this particular fraud. That, however, is no more than “but for” causation. There is no parallel between, on the one hand, the continuation of the loss-making trading businesses on the strength of dishonest statements to the auditors and, on the other hand, dishonest misappropriation of company funds, other than the presence of dishonesty. The absence of any such transaction, or any other misappropriation, in FY2009 meant that there was no reason to anticipate the transaction. There was no effective causal link between the transaction and GT’s negligence in the 2009 audit. I have earlier set out the passage from the judgment at [1001] where the judge held GT liable for this loss on the basis that it was “part of the management’s dishonest trading” and that it occurred “through AssetCo’s (continued) trading in a fundamentally dishonest manner”. In this paragraph, the judge is relying on the Jaras transaction as part of the fraudulent conduct of the business, but for the reasons earlier given, that is not a basis available on the common ground between the parties.

Although legal causation was accepted by the parties and the judge as a distinct element in establishing the claim against GT, virtually all the arguments that might be deployed on it formed part of the submissions on the scope of GT’s duty.

Reflecting the parties’ submissions, the judge dealt with legal causation largely as part of scope of duty. He did not conflate the two but at [972] he observed, correctly in this case, that the two were closely connected and once the recoverability of the trading losses had been dealt with, there were only two remaining issues, reliance and novus actus interveniens. He resolved those issues in favour of AssetCo and there is no appeal against his conclusions.

The submissions before us on legal causation, separately from the scope of GT’s duty, were very short indeed. In its skeleton argument, GT restated without repeating its submissions on the judge’s error in proceeding on the basis that AssetCo had been run in a fundamentally dishonest way and that GT had a duty to detect it. In his oral submissions, Mr Salzedo said that legal causation could not be satisfied without a finding as to the causes of the losses suffered by AssetCo. In my judgment, as the judge found, the losses occurred because an insolvent business was allowed to continue when it was only “ostensibly sustainable” by reason of the failure of GT to detect the dishonesty of the audit representations made to it by management, such failure depriving AssetCo of the opportunity to which it was entitled of correcting the position. As Mr Salzedo said, the issue has been presented as essentially the same as scope of duty.

It was common ground by the time of the appeal that, if GT had conducted a competent audit, the board would have known that AssetCo was insolvent by the end of May or early June 2009 at the latest. It was also common ground that, without a scheme of arrangement with its banking and other creditors and a restructuring of its share capital, AssetCo would have gone into insolvent liquidation. It was therefore critical to AssetCo’s case that it lost a real and substantial opportunity to put in place such a scheme and restructuring.

The judge concluded that AssetCo would have successfully completed a scheme and restructuring in 2009. He found that AssetCo had established that it would have taken all the steps necessary to achieve this and that the chances of third parties doing what was necessary for this purpose were in each case either 100% or so high that they fell to be treated as 100%. He accordingly did not discount the amount awarded as damages to take account of any chance that the scheme and restructuring would not have been put in place.

GT challenges this conclusion on two principal grounds. First, even on his own findings, the judge was wrong to treat each third party contingency as a certainty. While in some cases he specifically assessed the relevant chance at 100%, in others he assessed them at, say, “not less than 90%”. He was wrong to round those chances up to 100% but should have treated them as 90% and multiplied the percentage chances of each contingency in order to find the overall chance of a successful scheme and restructuring in 2009 and awarded damages on that basis. Second, GT challenges the percentage prospects found by the judge in the cases of four specific contingencies, submitting he should have assessed the chances at a much lower level. When multiplied with the other percentages, this would of course produce a significantly lower overall chance and would therefore significantly reduce the quantum of damages.

In order to put these challenges in context, it is necessary to summarise the steps in the judge’s analysis of the Counterfactual. He addressed the legal and evidential issues in considerable detail in Section G (paragraphs [351]-[877]) of his judgment.

The judge addressed the legal issues in section G.2 (paragraphs [361]-[460]).

Having reviewed authorities from Allied Maples Group Ltd v Simmons & Simmons [1995] 1 WLR 1602 to Wellesley Partners LLP v Withers LLP [2016] Ch 529, the judge at [415] held that where the loss claimed depends on a hypothetical course, the claimant must prove on the balance of probabilities what it would have done but, in order to satisfy the requirements of causation, need only show that there was a real or substantial chance of any necessary action by a third party. The evaluation of that chance was part of the assessment of the amount of the damages. He rejected the submission of AssetCo that where a claimant is able to prove on the balance of probabilities that the defendant’s breach has caused it to suffer loss, it is entitled to recover damages equal to the full amount of such loss, even where the claim depends on the hypothetical acts of third parties. This is not challenged by AssetCo. In such a case, while causation will be established and, for the purposes of quantification also, the claimant will have established what it would have done, the court must still evaluate the chance that the hypothetical would have occurred, essentially evaluating the chances that the relevant third parties would have taken the necessary action.

At [416], the judge observed that in a particular case the court may have before it evidence which allows it to conclude that the hypothetical conduct of a third party is, adopting the words of counsel for GT, “a racing certainty…so close to 100[%] that the judge awards 100[%]”.

AssetCo’s counterfactual case in 2009 depended on a number of steps which the judge summarised at [353] as follows:

The availability of Mr Davies to be appointed by NAV to carry out further investigations, and then as executive chairman of AssetCo;

An agreement between NAV and the Board of AssetCo, pursuant to which NAV pledges support for AssetCo and Mr Davies is appointed as executive chairman by 8 June 2009;

The resolution of Shannon’s and Flynn’s roles after Mr Davies’ appointment;

Mr Davies’ other actions once appointed as executive chairman of AssetCo;

The reaching of standstill agreements with the London and Lincoln banks;

Staving off insolvency in the period leading up to the proposed scheme;

Support of AssetCo’s shareholders for the proposed scheme and placing;

Support of AssetCo’s creditors for the proposed scheme;

Conclusion of the SOC Contract after the scheme.”

In turn, these steps depended on the conduct and actions of various third parties: NAV or Mr Mills; Mr Shannon and Mr Flynn; AssetCo group companies; banks; and other creditors.

The judge recorded at [358] GT’s submission that “in relation to every step which depends upon the conduct of a third party the court must assess the likelihood of such step being achieved expressed as a percentage, and discount any damages accordingly, multiplying such percentages together where there are additional steps involving third parties, thereby further reducing any damages recoverable”.

The judge returned to this point, and referred to another, at [417]:

In section G.2.5 of his judgment, the judge addressed the first of these issues. At [419] he quoted GT’s submission that: “Where there are a number of independent steps which would have needed to be taken by specific third parties before the claimant could have obtained a benefit, the probability of each step must be multiplied together”. GT relied on a number of authorities (Phillips & Co v Whatley [2008] Lloyd’s Rep IR 111, Harrison v Bloom Camillin (No 2) [2000] Lloyd’s Rep PN 89 and Joyce v Bowman Law Ltd [2010] EWHC 251 (Ch), [2010] PNLR 22) to support what the judge called “this (mathematical) approach”

At [421], the judge recorded that by the time of closings, it was common ground that this mathematical approach was only appropriate where the steps are truly independent or discrete in nature, as recognised by this court in Hanif v Middleweeks

From [432], the judge considered submissions made as regards the approach taken by

GT submitted that this approach was only appropriate where the contingencies were not independent or there was an overlap of the type identified in Hanif, while AssetCo submitted that Proudman J was correct in the approach she adopted. Otherwise, it submitted, there was a risk that it might recover very little even if it proved a strong prospect that each of the contingencies would occur. This would be all the more unjust in a case, such as the present, where the investigation was into what might in the past have happened, rather than a prediction of future events. AssetCo relied on the decision of Males J in Fiona Trust & Holding Corporation v Privalov (No 2) [2016] EWHC 2163 (Comm) [2017] 2 All ER 570, [2017] 2 All ER 570. The judge rejected AssetCo’s submission. Fiona Trust was a case where the claimant was seeking damages for lost profits, where an overall assessment is appropriate. Different considerations apply in claims based on the loss of a chance, particularly where there are independent contingencies that do not overlap: see [445]. To the extent that in Harding Homes the contingencies were truly independent, the judge did not agree with the approach taken in that case.

The judge also rejected AssetCo’s reliance on its point that a claimant might recover little in a case where there was a strong prospect that each contingency would materialise, because “if there are a number of independent steps which would need to be taken by specified parties before the claimant could have obtained the benefit then it is the product of those contingencies that reflects the loss”: [448].

At [456]-[460], the judge considered the second of the two points set out at [417], concluding that the evidence of the third party does not obviate the need to assess the chances that he would have acted as alleged, but “the reality is that having heard a witness the court is much better placed in its assessment of the chance, and indeed it may be so sure of what would have happened as a result of hearing from the witness that it can assess the chance at 100% (or so high as to be treated as 100%)”.

Referring to the steps in the Counterfactual, the judge noted at [486] that “not all the contingencies are necessarily truly independent (see Tom Hoskins), and factors affecting the cumulative future events overlap and/or are affected by the same consideration so that, as is common ground, an entirely mathematical approach is not appropriate (see Hanif)”. This would affect whether, and to what extent a purely mathematical, or percentage on percentage, approach was to be adopted.

At section G.4.1 (paragraphs [489]-[564]), the judge addressed the issue of the date(s) by which the material audit matters would have been uncovered by a competent auditor. His conclusion, accepting AssetCo’s case, was that the end of May 2009 was the latest time by which those matters would have been known to the audit committee

From [583], the judge considered in detail the chances of each of the other hypothetical third party acts.

The judge first considered whether Mr Davies would have been offered, and would have accepted, appointment as a director of AssetCo, which was itself linked to whether NAV would have pledged short-term support for AssetCo while Mr Davies investigated the position and attempted to secure a long-term solution by means of a scheme of arrangement. GT accepted that in principle NAV would have wanted to appoint Mr Davies to carry out investigations. The judge stated at [586] that “I am satisfied that not only is there a real and substantial chance, but it is an absolute certainty that once NAV (through Mr Mills) became aware of AssetCo’s financial problems Mr Mills, on NAV’s behalf, would have approached Mr Davies immediately by email and asked him to go on AssetCo’s board and carry out further investigations”. At [587]-[608], the judge considered whether Mr Davies would have accepted the offer of appointment, concluding as regards both Mr Mills asking Mr Davies to go on to the board and Mr Davies agreeing to do so, that “there is no doubt in my mind that that is what they each would have done, and as such the probability …does not stand to be discounted in relation to this step”.

The next, but related, question was whether NAV would have pledged short-term financial support for AssetCo in return for the appointment of Mr Davies as executive chairman. Substantial submissions were made to the judge by GT as to why the chances of such support were not high, which were considered by the judge at [609][672]. He concluded at [671] that “there is no doubt in my mind that…NAV would have so agreed, and would have supported AssetCo, and the possibility of it not doing so, and instead taking legal action against AssetCo, can be dismissed, such that the likelihood of such step is assessed as greater than 90% (and so as 100%)”. The judge was “certain that Mr Davies would have [taken on the role of executive chairman], and as such I assess the likelihood of such step as 100%”.

The judge then considered the subsequent third party actions necessary for the Counterfactual, starting with the chances of agreeing a standstill with Lloyds Bank and the other banks which had outstanding loans to AssetCo London and AssetCo

The resolution of the roles of Mr Shannon and Mr Flynn is dealt with at [725]-[729]. The main significance of this lies in the attitude that they would have taken to the scheme and restructuring, which the judge considered later in his judgment. In this section, the judge concluded that they would have ceased to be directors but would have remained as employees for as long as AssetCo considered that necessary, partly

The judge addressed the financing of AssetCo in the period from Mr Davies’ appointment as executive chairman to approval of a scheme at [705]-[724] and [730][800].

AssetCo’s case was that, with Mr Davies’ appointment, it would have ceased all nonessential expenditure. This would have included all expenditure for loss-making subsidiaries, in particular AssetCo London and AssetCo Lincoln. They would have become self-financing, with their banks agreeing to the use of their cashflow under their fire services contracts to remain in business, as happened in 2011. The only expenses that AssetCo would have needed to meet were its own operating costs and the professional and other costs involved in the preparation of the scheme of arrangement, amounting in total (as was agreed) to some £2.4 million. This included about £1.46 million in respect of the scheme, which for the most part would not need to be paid until the scheme was approved. AssetCo’s needs in the meantime would have been met by funds provided by NAV.

GT’s case was that AssetCo would have gone into liquidation, or there was a material chance that it would do so, before a scheme could be approved. It pointed to the insolvent position of AssetCo and the London and Lincoln companies and submitted that, among the many creditors who would not be paid in this period, there would be some who would seek to wind up AssetCo.

The judge held that the burden lay on AssetCo to prove its survival through to the approval of a scheme and found that it discharged the burden. He examined in detail the evidence as regards each group of creditors and concluded that none of them would have sought to wind up AssetCo, not least because it would make no commercial sense for them to do so. As against a return of between nil and 0.54 pence in the pound on a liquidation, they would know that they stood to receive substantially more under a scheme; under the scheme in 2011, the return was 23 pence in the pound. He had “no doubt” that NAV would have supported AssetCo during the period needed to investigate AssetCo’s position and propose the scheme. He found that Mr Mills was confident that AssetCo would be awarded business in Abu Dhabi, which he viewed as highly profitable. No discount was required “to take account of any (hypothetical) action of trade creditors” and the evaluation of NAV supporting AssetCo in this period “does not stand to be discounted, and stands at 100%”.

The judge addressed the support of shareholders and of creditors for the scheme at

The judge’s overall conclusion at [873] was that AssetCo would have entered into the scheme in 2009 and that “at the quantum stage the chances of AssetCo doing so do not stand to be discounted in relation to any of the necessary steps en-route”. He added that this was perhaps unsurprising, given the successful scheme in 2011.

The last in the list of steps in the Counterfactual involving third parties was entering into the SOC contract after the scheme. The judge’s conclusion that the contract would have been made is also challenged by GT, and again I will consider the judge’s reasons when addressing that challenge.

Accordingly, the judge found that AssetCo had established its case on the Counterfactual and that no discount fell to be made to the amount of damages by reference to loss of chance principles.

Before coming to the overall issue as to whether the judge should have adopted a mathematical approach to his assessment of the chances of the Counterfactual occurring, it is logical first to consider GT’s challenge to his findings on four individual steps involved in the Counterfactual: the support of NAV for AssetCo, the support of Mr Shannon and Mr Flynn for the scheme, the support of group creditors for the scheme, and the conclusion of the SOC Contract.

The correct approach for an appeal court to adopt to the review of findings as to hypothetical facts was debated before us.

Mr Templeman for AssetCo submitted that the same approach was to be adopted to such findings as applied to other findings of primary fact. That approach has been authoritatively set out in McGraddie v McGraddie [2013] UKSC 58, [2013] 1 WLR 1911 and other cases. As Lord Hodge expressed it in Henderson v Foxworth Investments Ltd [2014] UKSC 41, [2014] 1 WLR 2600 at [67]:

Mr Salzedo submitted that the principles reiterated in cases such as McGraddie v McGraddie applied to findings of fact as to what has occurred, made in the usual way on the balance of probabilities. The assessment of probabilities on hypothetical facts is a different exercise, calling for a different approach by the judge and by an appellate court. The judge is not required to decide on a binary basis whether something in fact occurred, but must assess the relative chances whether something would and would not have occurred. While an appellate court will give “some latitude” to the judge’s assessment, it will more readily interfere than is the case with findings of fact. He relied on the decision of the Supreme Court in Re B (A Child)

Mr Salzedo is right to draw a distinction between a finding of fact on the balance of probabilities and an assessment of the chances of a hypothetical fact. This distinction was clearly drawn by the Supreme Court in Perry v Raleys Solicitors [2019] UKSC 5, [2020] AC 352 where Lord Briggs, with whom the other members of the Court agreed, held that in a loss of a chance case the judge was right to hold “a trial within a trial” of those matters which the claimant had to establish on a balance of probabilities, i.e. what the claimant would have done in a counterfactual, whereas this was not generally appropriate to the assessment of a chance, i.e. of what others would have done or the occurrence of an extraneous event: see [20] and [31].

The assessment of a chance is, as Mr Salzedo accepted and Lord Briggs said in Perry v Raley Solicitors, an evaluation. The extent to which an appeal court may interfere with an evaluation is likely to differ according to the subject of the evaluation. Some evaluations involve mixed questions of fact and law, for example whether a contract is a contract of employment or a contract for services or whether an invention is obvious for the purposes of a patent. These are evaluations based on findings of primary fact. Even in cases where the findings of primary fact are not challenged, the authorities establish that an appeal court should take a cautious approach to interfering with the judge’s decision. In Datec Electronics Holdings Ltd v United Parcels Service Ltd [2007] UKHL 23, [2007] 1WLR 1325 at [46], Lord Mance (with whom the other members of the House agreed) cited with approval paragraphs [16] and [17] of Clarke LJ’s judgment in Assicurazioni Generali SpA v Arab Insurance Group [2002] EWCA Civ 1642, [2003] 1 WLR 577. Clarke LJ said at [16]:

At [17], Clarke LJ cited from the judgment of Mance LJ in Todd v Adam [2002] EWCA Civ 509, [2002] 2 All ER (Comm) 1 where Mance LJ said at [129] that “so far as the appeal raises issues of judgment on unchallenged primary findings and inferences, this court ought not to interfere unless it is satisfied that the judge’s conclusion lay outside the bounds within which reasonable disagreement is possible”.

The approach to evaluative judgments or assessments has been further considered in cases since Re B (A Child). In R (on the application of AR) v Chief Constable of Greater Manchester Police [2018] UKSC 47, [2018] 1 WLR 4079, the Supreme Court was concerned with the correct appellate approach to the judge’s assessment of proportionality in public law of an administrative decision, raising issues under articles 6 and 8 of the European Convention on Human Rights. Giving the judgment of the court, Lord Carnwath said at [64]:

This approach was applied by this court, when considering an appeal against the trial judge’s decision that the exclusion of a director from participation in the management of a company was not justified by his conduct and that it was, in the circumstances of the case, “unfairly prejudicial” conduct for the purposes of section 994 of the

While the evaluations undertaken by the judge in the present case are not findings of fact on the balance of probabilities, they are conclusions reached on the basis of a close examination of the relevant evidence, some oral and, perhaps to a greater extent, some documentary. The judge was immersed in this material over an extended period and also received very lengthy and detailed written and oral submissions from the parties. His evaluation did not depend on any principles of law but was directed to very specific and fact-sensitive hypothetical issues, unique to this case. Leaving aside the oral evidence, many of the factors that are now recognised as reasons why appeal courts should be very slow to interfere with findings of fact also apply here: see, for example, McGraddie v McGraddie per Lord Reed at [3]-[4]. These are among the factors to which Lewison LJ referred in a well-known part of his judgment in Fage UK Ltd v Chobani UK Ltd [2014] EWCA Civ 5, [2014] ETMR 26. As he said at [114]: “Appellate courts have been repeatedly warned, by recent cases at the highest level, not to interfere with findings of fact by trial judges, unless compelled to do so. This applies not only to findings of primary facts, but also to the evaluation of those facts and to inferences to be drawn from them”.

In my judgment, in deciding whether the judge was “wrong” in any of his evaluative conclusions, this court should recognise not only that it was his role to decide these issues but also that, in doing so, he enjoyed advantages not available to this court, and we should interfere only if his conclusions were not ones reasonably open to him on the totality of the evidence, unless there is an underlying flaw in his reasoning which means that his assessment cannot stand.

GT submits that the judge’s finding that NAV would have supported AssetCo

GT submits that the reasons given by the judge cannot support his conclusion. First, the prospect of an upside on the warrants to subscribe for shares in AssetCo was not cited by Mr Mills and is “incoherent”. The strike price was 61.2 pence per share which would not have been realistic in the Counterfactual. Second, the financial interests of AssetCo and NAV would have been aligned only if NAV provided financial support for AssetCo and cannot therefore provide a reason for NAV giving such support. I interpose here to say that this involves a misreading of what the judge said at [633]; it was a comment on the effect of support by NAV, not a ground for thinking that NAV would have given support. Third, the desire of the Abu Dhabi authorities “to contract with the mother ship” in February 2011 was after they had entered into the SOC Contract and before the true facts about AssetCo were known. It was of limited value in assessing NAV’s likely perception in June 2009. Again, I interpose to say that this does not properly reflect what the judge said at [634], his substantial point being that for other reasons NAV considered it important to have AssetCo in place when bidding for business in Abu Dhabi. The mother ship reference simply provided further support from Abu Dhabi for NAV’s view. Fourth, the judge did not properly address the “obvious problems” posed for NAV by Mr Shannon’s ability to block a restructuring as a result of holding shares carrying over 25% of the voting rights.

Finally, the judge’s finding that NAV would have supported AssetCo in 2009 despite not knowing the true scope of its liabilities, unlike the position in 2011, was inconsistent with Mr Mills’ own evidence that he would not invest without knowing or thinking he knew the true extent of liabilities. There was no finding that it would have been possible to ascertain the extent of the liabilities that would have to be met before a scheme was approved in 2009. NAV would not therefore have been able to limit its short-term support. This was a key difference between the situation in 2009 and in 2011 when NAV had been closely involved in AssetCo for some time, Mr Davies being appointed a director some six months before the scheme and after five weeks of dialogue. Further, AssetCo had raised capital of £16 million in March 2011 and it was not understood to be insolvent, although some creditors were pressing for payment.

GT submits that the judge should have assessed the chances of support by NAV at no more than 50%. He was wrong to prefer the evidence given by Mr Mills that NAV

The judge fully accepted Mr Mills’ evidence that NAV would have supported AssetCo through to a scheme. First, he accepted Mr Mills’ evidence that he believed (indeed, there was never any doubt in his mind) that highly profitable business could be won in Abu Dhabi and that in the summer of 2009 such business would be awarded “fairly imminently”. He accepted Mr Mills’ evidence that he believed that, if AssetCo were driven into insolvency, it would almost certainly mean that any prospect of business in Abu Dhabi would be lost. He regarded AssetCo’s experience of the type of operations under consideration as important to obtaining any business in Abu Dhabi. NAV’s interest was in pursuing profitable business in Abu Dhabi, not in cutting losses and seeking reimbursement of what remained of the funds it had invested in AADL. As Mr Mills put it in evidence, “it was Abu Dhabi that was going to make us rich”.

Second, by agreeing to provide financial support, NAV would not have exposed itself to an unlimited liability. There would have been no question of NAV underwriting all of AssetCo’s liabilities. In the first instance, it would have provided short-term support in return for Mr Davies’ appointment as chairman, following which liabilities would have been investigated in the context of proposals for a long-term solution. The judge accepted Mr Mills’ evidence that, where the payment of any debt could not be avoided, NAV would have consented to the use of the PSA monies for that purpose,

I will consider the position of Mr Shannon later, but for the reasons given there the judge was not required to weigh it as a factor that would or might realistically have deterred NAV from providing short-term support pending a scheme.

The judge assessed Mr Mills’ evidence against the various factors relevant to whether NAV would have supported AssetCo through to a scheme and there was ample evidence for his conclusions.

GT next challenges the judge’s assessment of the chances that Mr Shannon and Mr Flynn would have opposed the restructuring of AssetCo’s capital which accompanied the scheme. The scheme was proposed between the company and its creditors, but it was dependent on the capital restructuring to implement the debt for equity swap for which it provided. This required the passing of special resolutions, for which a majority of 75% of the votes cast was needed. As Mr Shannon in 2009 held shares carrying 37.5% of the votes, he could block a special resolution. Mr Flynn’s holding gave him only 9.85% of the votes and his position does not therefore need separate consideration. By contrast, in 2011 the combined holdings of Mr Shannon and Mr Flynn carried less than 25% of the votes, as a result of the capital raising earlier that year in which they did not participate.

AssetCo had to satisfy the judge that there was a real and substantial chance that the special resolutions would have been passed and therefore that Mr Shannon would not have voted against them. If so satisfied, it was then for the judge to assess the chances that he would not have done so.

The judge dealt with this at [840]-[859]. AssetCo relied on Mr Shannon’s close relationship with Mr Mills and on Mr Shannon’s knowledge that his shares would be worthless in the absence of a solvent solution. As against this, GT pointed to his conduct in 2011 when Mr Shannon took steps contrary to the interests of AssetCo, although in the end he did not vote against the resolutions or take steps to oppose the scheme or restructuring. In 2009, when he had a blocking vote, GT submitted that he would have appreciated his bargaining power and used it to achieve advantages for himself. GT submitted that there was no real or substantial possibility that Mr Shannon would have supported the restructuring.

While taking account of Mr Shannon’s actions and ultimate inaction in 2011 and the benefit to him of a solvent solution, the judge identified “a very good reason, and indeed a reason which I consider would have been the determining consideration for him, namely that without a solvent solution for AssetCo he would, in all likelihood, have been at risk of bankrupting himself by forcing AssetCo into insolvent liquidation”. Mr Shannon had pledged 8.8 million of his shares to a bank as collateral for a loan facility of £2.7 million. From January 2009, Mr Shannon was coming under increasing pressure from the bank to repay or refinance the facility. The judge considered that it was “far-fetched to suggest that Mr Shannon would ‘cut off his nose to spite his face’ in the situation he faced at that time”. A solvent solution was obviously in his best interests and, consistent with the belief of Mr Mills and Mr Davies, “I have no doubt that he would have approved (or not opposed) the various resolutions”. The judge also took account of the fact that Mr Shannon had less to cover up in the way of his own misfeasance as a director in 2009 than in 2011 and therefore had less to fear from AssetCo continuing to operate. The chances of Mr Shannon voting against the resolutions were “vanishingly small, and not such as would justify any discount of any damages recoverable”.

GT challenge this finding. The view that Mr Shannon would not cut off his nose to spite his face is undermined by the fact that under the capital restructuring the holdings of existing shareholders would have been virtually wiped out. On the same basis as the 2011 restructuring, shareholders would have received 1 new share for 1,000 existing shares. Mr Shannon would have held 26,963 shares in place of over 29.63 million shares, while 3.75 million new shares would have been issued under the scheme. His resulting shareholding would have been so small that it would not have saved him from his financial difficulties. The judge should have found that there was no more than a 50% chance of the special resolutions being passed.

GT’s challenge really comes to saying that, because Mr Shannon behaved for a while in 2011 in a fashion that Mr Salzedo rightly characterised as “quixotic and very aggressive”, there is a material chance, even a 50% chance, that he would likewise have tilted at windmills in 2009.

In my judgment, the judge was entitled on the evidence to dismiss this chance as so small as not to justify any discount.

I reach this view for several reasons. First, opposition by Mr Shannon to the rescue of AssetCo would have been wholly irrational. While it is fair to say that his holding after the scheme would have had a very small value, something is better than nothing, especially if the holder is facing possible bankruptcy and the shares or some of them are charged in favour of his principal creditor. Wilfully destroying such value as remained in the security is not likely to have been well received by the creditor. Indeed, the creditor would have been entitled to take steps to preserve its security.

Second, GT’s thesis is that Mr Shannon would have held out for some inducement, over and above what was offered to other shareholders, to secure his support. There is, however, no evidence to suggest that AssetCo or NAV would have succumbed to commercial blackmail of this sort. GT was not specific, so far as I am aware, as to the form that this inducement would have taken, but if indeed it were provided to secure his support, it would of course have to be disclosed to the other shareholders. It underlines how unlikely it would have been that such an approach by Mr Shannon would have succeeded.

Third, the judge found that, although Mr Shannon would immediately have left the board in the Counterfactual, he would have been retained by AssetCo in some capacity. The judge said at [728]:

Fourth, Mr Shannon had a close relationship with Mr Mills and the judge found that it would have been Mr Shannon himself who approached NAV for support and that he would have supported the appointment of Mr Davies as chairman in return for its support.

Accordingly, I do not consider that there is any basis on which this court could properly interfere with the judge’s assessment as to Mr Shannon’s conduct in the Counterfactual.

The third finding specifically challenged by GT, which it put at the forefront of these challenges, was that creditors would not have supported a scheme in 2009, the chances of them not doing so being less than 10%.

The scheme could succeed only if it was approved, as required by section 899 of the Companies Act 2006, by a majority in number representing three quarters in value of those creditors voting at the meeting, or (if more than one class) at each meeting, of creditors. Substantial sums were owed by AssetCo to its subsidiaries. In particular, it owed some £65 million to AssetCo Emergency Limited (AssetCo Emergency), representing approximately 65% of the amounts owed to all group creditors and some

Under the terms of the scheme, group creditors were to receive a return of 0.01p in the pound, whereas non-group creditors were to receive 23p in the pound. On this basis, AssetCo Emergency would receive approximately £6,500 in respect of its debt. In the only realistic alternative of an insolvent liquidation, it was estimated in the circular to creditors in 2011 that the return to creditors was likely to be between nil and 0.54p in the pound, the precise amount depending principally on the level of liquidation costs. Assuming, unrealistically, that the return would be 0.54p in the pound, AssetCo Emergency would receive approximately £350,000.

GT submitted to the judge, and submits to us, that group creditors could not lawfully vote in favour of a scheme on these terms in 2009. To do so would be a breach of duty by the directors of those companies. It could not be in the interests of those companies to receive so much less than the amount offered to non-group creditors and, more importantly, less than the likely return in a liquidation. Counsel for GT accepted at the trial that a return of about half a penny in the pound is not much, but it was 50 times better than 0.01p in the pound. With AssetCo, in the Counterfactual, no longer supporting its subsidiaries, there was no reason for AssetCo Emergency to accept a scheme on these terms.

The judge relied on the fact that in 2011 the group companies supported a scheme on precisely these terms “and did so unanimously in circumstances where it is to be assumed that their directors acted in accordance with their duties, and were properly legally advised (there being no evidence to the contrary)”. He regarded the actual events of 2011 as being the best evidence of what would have occurred in 2009.

The judge supported that conclusion by reference to objective considerations. It was inherently unlikely that a subsidiary would destroy its parent company, absent truly compelling reasons for doing so, thereby exposing its own future to substantial uncertainty. As a result of the scheme, the subsidiaries would have a parent company with every prospect of being successful for itself and the group in the future. The directors of subsidiaries could reasonably decide that this, rather than “taking a chance” in a liquidation (as it had been put on behalf of GT), was in the best interests of their companies. When it was put to Mr Davies that the directors would have decided to take a chance with a liquidation, he said “I do not think any people inside a group would do something like that”. The judge found this evidence to be entirely credible.

GT submits that the judge failed to make any findings as to the legal advice that the directors of the group companies would have received nor did he find how, lawfully, they could have supported a scheme in 2009. The burden of establishing this lay on AssetCo and it failed to discharge it. It could not rely on what occurred in 2011 to establish that the same could lawfully occurred in 2009. There was no evidence to support the judge’s “speculation” that the directors would have considered there to be any commercial benefit in continuing as subsidiaries of a solvent parent, when they were required to sink or swim on their own. As for Mr Davies’ evidence, he was not involved in 2009 and he had no special expertise.

In my judgment, not only was the judge entitled to look to the events of 2011 to determine what would have occurred in 2009, it was the obvious thing to do. It was not suggested that the directors acted in breach of duty in causing subsidiaries to vote in favour of the scheme in 2011 and, in the absence of a challenge, it is not to be assumed that they did so. Nor is it suggested that there were any differences between the circumstances in 2009 and 2011 material to this issue. No reason is put forward why the directors should have taken a different view in 2009 than the view they took in 2011. On that basis, as it seems to me, AssetCo discharged the burden of establishing that group companies would have supported a scheme in 2009 and the judge was entitled to reach his conclusions on this issue.

The judge was also entitled to consider that the objective factors to which he referred supported his finding. There is nothing implausible – in fact, the reverse – in the directors of subsidiaries taking the view that the interests of their companies are best served by a solvent parent company, rather than a parent company in insolvent liquidation. Moreover, counsel for GT were unable to tell us whether, for example, AssetCo Emergency, which was simply an intermediate holding company, itself had any creditors whose interests the directors would have been required to consider. If it had no creditors, the directors would properly be guided by the interests of its parent company. It is also the case that the proposals announced in 2011, which it is assumed would have been mirrored in 2009, included arrangements for the London Group; see the judgment at [343]. As the holding company of the London Group, this would be an important consideration for the directors of AssetCo Emergency in deciding whether to support the scheme.

I therefore reject GT’s challenge to the judge’s conclusion as regards the support of group creditors, including his finding that the chances of them not supporting a scheme were less than 10%.

The fourth and final specific assessment challenged by GT relates to the chances of concluding a contract with SOC, assuming that a scheme was approved in 2009. Since there is no question of an SOC contract being concluded in advance of the approval of a scheme, this is a rather different contingency to the others. GT’s submission to the judge was that, rather than winning a contract in Abu Dhabi, AssetCo would not have avoided insolvency even if a scheme had been approved in 2009: see the judgment at [806]. AssetCo’s case was that, even without an SOC contract, AssetCo would have remained solvent after the approval of a scheme. The essential issue was not whether AssetCo would have concluded an SOC contract, but whether it would have remained solvent.

It is for this reason that when the judge, having reviewed the evidence, said at [828] that he was satisfied that AssetCo would have won the SOC Contract and another contract in Abu Dhabi, he added “should that be relevant to any of the issues that arise”. He repeated the point at [836].

The judge considered the prospects for AssetCo’s solvency after approval of a scheme in the Counterfactual at [837]:

I did not understand GT to challenge this finding, but rather to challenge the judge’s finding that AssetCo would have won a contract with SOC after a scheme in 2009 had been approved. It submits that it was not a finding open to him on the evidence and that he should have found that there was no real or substantial chance that it would have concluded an SOC contract, or alternatively that the chances of doing so were very low indeed. I have reviewed the reasons given by the judge and the submissions made by GT, concluding that there are no grounds for this court to interfere with the judge’s finding. It is not, however, necessary to go further because, in the light of the findings of the judge at [837], it is not a point that assists GT.

For the reasons given above, I reject the challenges made by GT to the judge’s conclusion on each of the four specific issues.

I turn therefore to consider GT’s case that the judge should have adopted the mathematical approach of multiplying the probabilities in the case of all contingencies where he did not hold the chance to be 100%. GT accepted that this approach was not appropriate where the contingencies were not independent. As applied to the present case, GT accepted that steps 1 to 5 in the list set out by the judge at [353] were not independent. Step 1, the hypothetical discovery of AssetCo’s major problems in 2009 if GT had not been negligent, was admitted by GT and is not so much a contingency in the hypothetical as its starting point. Steps 2 to 5 all concern NAV, Mr Mills and Mr Davies. These steps were so closely linked that GT was content to treat them as a single contingency. In fact, there is little content to step 5, because its main elements are comprised in steps 3, 4 and 7, and it does not require separate examination.

GT submitted that steps 6 to 9 were properly regarded as four separate contingencies.

The judge did not reach any conclusion on the independence of the individual contingencies. In an overall commercial situation such as the 2009 Counterfactual, there is some dependence between all these steps. Nonetheless, I would accept that there is a sufficient independence for these purposes between steps 2 to 5 and 7 (forming a single contingency) and step 6, step 8 and step 9, to make four independent contingencies. Step 10 is not a relevant contingency, for the reasons previously given.

While the judge accepted that it was generally right to adopt a mathematical approach, by multiplying the chances of each of independent contingency, he did not do so on account of his very high assessment of the chances of each contingency.

GT submits that the judge was wrong not to adopt the mathematical approach. Although GT accepts that a judge is entitled to find that a contingency was a certainty, it says that this is not what the judge did in this case. Rather, it submits, the judge proceeded on a misunderstanding of the submissions made to him on behalf of GT, believing incorrectly that GT accepted that, if a contingency had a 90% chance, it should be treated as a certainty. In submissions to the judge, counsel had accepted, indeed put forward, that in two circumstances a contingency with a 90% chance could be treated as a certainty. First, where the loss of a chance depended on a single contingency with a 90% probability, the contingency and hence the lost chance could be treated as a certainty. Second, where the loss of a chance depended on more than one contingency, and the result of multiplying the contingencies produced an overall chance of 90%, the lost chance could be treated as a certainty. However, GT submits that the judge applied the same principle to the individual contingencies in this case but did so in error because the alleged loss of a chance in this case depended on more than one independent contingency.

There was argument before us as to whether GT had accepted or conceded that this was the right approach and whether the judge understood (or misunderstood) GT to have made this concession. We were taken to passages in the closing submissions of counsel for GT and to submissions made to the judge on this subject after circulation of his judgment in draft and his responses to such submissions. I have re-read the relevant submissions and, although I think there could be reasonable confusion about it, I conclude that such a concession was not made by counsel for GT. It appears that the judge may well have understood such a concession to have been made. However, the real issue is the basis on which the judge reached his decision that the chances of the Counterfactual were 100% and that no discount should be made to this assessment. For this purpose, in my view, we should focus on the reasons given by him in his judgment. They are the reasons given by him for the order which is the subject of this appeal.

The judge’s overall conclusion at [873] was that “the chances of AssetCo [entering into the scheme in 2009] do not stand to be discounted in relation to any of the necessary steps en-route, for the reasons I have given” and that “no discount on the quantum recoverable stands to be made applying loss of chance principles in the light of the factual findings I have made in that regard”. It is therefore necessary to focus on the reasons he gave as regards each of the steps and “the factual findings I have made in that regard”. He used the events of 2011 as a cross-check, saying at [874] that his conclusion “is perhaps unsurprising given that AssetCo’s 2009 and 2010 Counterfactuals rely heavily on what actually happened in 2011”, finding that “for the reasons I have given, the same outcome would have occurred in 2009 or 2010 having

GT submits that as the judge found in a number of instances that the chances were in excess of 90% but declined to put a precise percentage on his assessment, when requested to do so after circulation of his judgment in draft, he must be taken to have assessed those contingencies at 90% and then applied a rule that chances of 90% are to be rounded up and treated as 100% chances.

I do not consider that this submission accurately reflects the process in fact adopted by the judge.

Taking first steps 2 to 5 and 7, the judge said that it was “an absolute certainty” that Mr Mills would have asked Mr Davies to become a director of AssetCo and that “there is no doubt in my mind” that he would have accepted (step 2). As regards the provision of financial support by NAV for AssetCo in return for Mr Davies’ appointment (step 3), the judge said that “there is no doubt in my mind that…NAV would have so agreed”, adding that the possibility of it not doing so can be dismissed, “such that the likelihood of such step is assessed as greater than 90% (and so as 100%)”. The judge was satisfied that Mr Shannon and Mr Flynn would have been retained as employees for as long as AssetCo considered that necessary (step 4). Staving off insolvency pending a scheme (step 7) is directly linked to NAV’s agreement to support AssetCo until a scheme was approved. The judge again said that there was no doubt in his mind that NAV would have provided whatever funding was needed for this purpose and the evaluation of the chance of NAV doing so does not stand to be discounted “and stands to be assessed at 100%”. Taking those steps together, it is clear that the judge concluded that it was a certainty that they would all have occurred. The reference to the chance of support from NAV being assessed as greater than 90% and so as 100% is to be read with the rest of what the judge said about this chance, namely that there was no doubt in his mind that NAV would have agreed to provide support.

Step 6 was reaching standstill agreements with the banks financing AssetCo London and AssetCo Lincoln. The judge said that it was “almost certain” that the largest lender, Lloyds Bank, would have agreed and that “the overwhelming likelihood is that the other banks would have followed suit based on Lloyds Bank’s stance, but again I cannot be certain. There is a very small chance that agreement would not have been reached. Looking at all the banks in the round…I assess the likelihood of this step being achieved as greater than 90% (and in consequence no discount is appropriate)”.

Step 8 was the support of shareholders for the capital restructuring which was essential to the scheme. In practice, the issue was whether Mr Shannon would have opposed it. The judge held that the chances that he would do so were “vanishingly small, and not such as would justify any discount of any damages recoverable”.

Step 9 was the support of creditors for the scheme, the issue being whether group creditors would have voted for it. The judge’s assessment was that “they would have so voted as evidenced by what each and every one of them did in 2011, the chance of them not doing so being less than 10%, and so not discounted”.

Leaving aside NAV’s support, the only steps about which the judge expressed himself in terms other than complete certainty were steps 6 and 9. In both cases, he assessed the chances at more than 90% and said that they did not stand to be discounted. The chances were not assessed as being 90%, but as being greater than 90%. The judge declined to be more precise, and in my judgment, it is understandable that he did so. Once a judge has assessed a chance as greater than 90%, an assessment that the chance is 93% or 96% or 99% lends a spurious degree of precision. It is not possible to assess the chances of the sort of events involved in the 2009 Counterfactual with such fineness. They are not events that can be tested in laboratory conditions. Having reached an assessment of greater than 90% for each contingency, he was, in my judgment, entitled in this case to treat it as being, in counsel for GT’s phrase, a “racing certainty”. This was not rounding up a 90% chance to a certainty but a conclusion that, within the confines of judicial decision-making, it was a certainty.

When this court in Allied Maple established that damages could be awarded on the basis of a loss of a chance, it recognised that such a degree of precision could not be achieved. Stuart-Smith LJ at pp.1613-1614 quoted from the speech of Lord Reid in Davies v Taylor [1974] AC 207 at 213 where Lord Reid said: “You can prove that a past event happened, but you cannot prove that a future event will happen and I do not think the law is so foolish as to suppose that you can. All that you can do is to evaluate the chance. Sometimes it is virtually 100 per cent certain: sometimes virtually nil. But often it is somewhere in between.”. Stuart-Smith LJ continued: “…the evaluation of the chance is part of the assessment of the quantum of damage, the range lying somewhere between something that just qualifies as real or substantial on the one hand and near certainty on the other. I do not think it is helpful to seek to lay down in percentage terms what the lower and upper ends of the bracket should be”.

In Hanif v Middleweeks at [40], Mance LJ assessed the chances of a particular contingency not occurring as being “reduced to the insignificant, in other words to a point where it could be altogether disregarded”. Giving the judgment of the Privy Council in Phillips & Co v Whatley [2007] UKPC 28, [2007] PNLR 27, in a negligence claim against solicitors involving three contingencies, Lord Mance said at [2] that the “conventional approach to a claim such as the present is…to measure its prospects of success and assess damages on a broad percentage basis”. The impossibility of reaching precision in a case such as this was recognised in Mr

In my judgment, the proper analysis of the judge’s reasoning is that he was satisfied that the chances of each contingency were so high that they fell to be regarded as certainties, not because of a principle or presumption that 90% equalled 100% but because a distinction between certain and almost certain was in this case meaningless. It was a conclusion that was open to the judge, both as a matter of principle and on the authorities.

This was an unusual case in that the Counterfactual in fact occurred, albeit two years later in 2011. The judge was right to attach significance to that, and to ask whether the differences in the circumstances between 2009 and 2011 were such as to make any material difference to the prospects for the Counterfactual. He was entitled to conclude that they did not do so.

I would therefore reject Ground 2 of GT’s grounds of appeal.

The judgment at [1227] lists the sources of funds used by AssetCo for the purposes of “its wasted expenditure on subsidiaries”, in other words making the inter-company loans to the subsidiaries which proved to be irrecoverable. Six sources are listed. GT contended that, if it was liable for the wasted expenditure, four of the sources represented benefits to AssetCo for which it must give credit.

The four sources in question include the proceeds of two share issues by AssetCo, the first in July 2009 raising just over £7.5 million and the second in March 2011 raising £16 million (of which £7,942,000 was used to fund wasted expenditure, but GT argues that the full proceeds represent a relevant benefit to AssetCo). The third source was amounts totalling £7,735,000 taken from the PSA monies and the fourth was a sum of £1,292,000 borrowed by AssetCo on overdraft. If credit is given for even some of these sums, it would eliminate the damages otherwise payable by GT.

In section J of his judgment ([1046] - [1090]), the judge rejected GT’s case that credit should be given for any of these amounts. GT cited well-known authorities on the compensatory purpose of an award of damages, such as Livingstone v Raywards Coal Co (1880) 5 App Cas 25 and Hodgson v Trapp [1989] AC 807 in which Lord Bridge said at 819: “If, in consequence of the injuries sustained, the plaintiff has enjoyed receipts to which he would not otherwise have been entitled, prima facie, those receipts are to be set against the aggregate of the plaintiff’s losses and expenses in arriving at the measure of his damages”. The judge characterised GT’s case as based on “but for” causation; the benefits would not have been available but for GT’s unqualified audit reports. Citing the Supreme Court’s decision in Fulton Shipping Inc of Panama v Globalia Business Travel SAU of Spain [2017] UKSC 43, [2017] 1 WLR

On that basis, the judge held that the claimed benefits did not arise out of, nor were they sufficiently closely connected to, GT’s breaches to satisfy the requirement for legal causation.

As regards the share issue in July 2009, the proximate or effective cause was

As regards the share issue in March 2011, it was too long after the audit of the 2009 accounts and there was no sufficient connection in fact or in time for GT’s breaches to be its legal cause. As far as the audit of the 2010 accounts was concerned, the proximate or effective cause of AssetCo’s decision to undertake the share issue was its need for cash, and (as the evidence showed) investors did not rely on GT’s audit reports.

AssetCo submitted to the judge that there was an alternative reason why credit for the proceeds of the share issues should not be allowed even if they were legally caused by GT’s breach: to do so would not be in accordance with justice, fairness and public policy. As The New Flamenco showed, this is established as a basis on which credit for a benefit may be denied. The reason advanced by AssetCo was that the investors would, on the authority of Caparo, have no claim against GT. If therefore AssetCo were obliged to give credit for the share issue proceeds, GT would to that extent have no liability to the company for losses for which it was otherwise liable but would also have no liability to the investors who had lost their investment. The judge said at [1086] that it was unnecessary for him to fall back on this, because he had held that the necessary legal causation was not established, but he considered AssetCo’s submission to be persuasive “and they would provide an alternative reason why credit need not be given if, contrary to my findings, legal causation had been demonstrated”.

More briefly, the judge also dismissed GT’s claim for credit in respect of the PSA monies and funds borrowed on overdraft into account.

GT challenges both the judge’s analysis of the applicable legal principles and, in any event, his application of those principles to the facts of the case.

As regards the legal principles, GT in substance repeats the submissions made to the judge. As they centre on the three recent Supreme Court decisions, I shall briefly summarise them.

Two of the cases, Swynson and The New Flamenco, were argued at the same time before the same panel, in November 2016. Judgment was given first in Swynson. The claimant made loans in reliance on a negligent due diligence report by the defendant accountants. Subsequently, the loans were repaid out of funds advanced for that purpose by the claimant’s owner. A number of issues were raised but, so far as relevant for the present appeal, the claimant sought damages from the accountant, arguing that account need not be taken of the repayment of its loans. Relying on authorities such as Parry v Cleaver [1970] AC 1, it argued that the refinancing loan made by the claimant’s owner was a collateral benefit not affecting its loss. This was unanimously rejected by the Supreme Court. The claimant’s loss was made good by the repayment of its loans; it had no loss to recover. The fact that the repayment was made out of funds advanced to the borrower by the claimant’s owner was irrelevant. The claimant and its owner were separate persons and the fact that the original loans

Judgment in The New Flamenco was given by Lord Clarke, with whom the other members of the court agreed. The owners of a vessel accepted a repudiatory breach of a time charter with an unexpired period of two years. There was then no available chartering market and they claimed damages for loss of profits during the remaining two years. They took the decision to sell the vessel, which realised a price of $23.7 million. The expert evidence established that its value at the end of the two-year period would have reduced to $7 million. The charterers claimed credit for the difference between these two amounts, on the basis that termination of the charterparty gave the owners the opportunity to sell at the higher price.

The Supreme Court held that the charterers were not entitled to this credit. A defendant was entitled to credit for a benefit secured by a claimant only if it was both factually (on a “but for” basis) and legally caused by the defendant’s breach. As Lord Clarke said at [30]:

On the facts of The New Flamenco, the owner’s interest in the capital value of the vessel had nothing to do with the interest injured by the repudiation of the charterparty. The vessel could have been sold at any time, including during the term of the charter provided the sale was subject to the charterer’s rights. The early termination of the charterparty did not necessitate a sale of the vessel. It was a decision taken by the owners on commercial grounds, independent of the termination. It was the absence of a relevant causal link that prevented the charterers from claiming this benefit. The analysis would be the same even if the owner’s decision to sell was driven by the lack of any work for the vessel. Lord Clark said at [33]: “At the most, that means that the premature termination is the occasion for selling the vessel. It is not the legal cause of it.”

Tiuta was heard and judgment was given after the other two cases, by a differently constituted panel except that Lord Sumption was a member of the panels for all three cases and gave the sole judgment in Tiuta. In that case, the claimant provided a loan facility for a term of nine months to a property developer. Shortly before the expiry of that term, in reliance on a valuation provided by the defendant surveyors, the claimant lent £3.088 million to the developer, on terms that just under £2.8 million would be used to repay the indebtedness under the first facility and the balance of £289,000 to finance completion of the development. The indebtedness under the first facility was duly discharged. None of the indebtedness under the second facility was repaid. The claimant alleged that the valuation was negligent and claimed damages equal to the unpaid amount due under the second facility. The defendant surveyors applied to strike out the claim, except to the extent of the “new money” of £289,000. They argued that the amount applied in repaying the first facility was not recoverable because if (as had to be assumed) the claimant would not have made the advances under the second facility but for the surveyors’ negligence, the advances under the first facility would have remained unpaid. That part of their loss would therefore have been suffered in any event.

The Supreme Court allowed an appeal in the surveyors’ favour. Lord Sumption said at

The claimant submitted that the discharge of the indebtedness under the first facility should be disregarded, because the application of the funds advanced under the second facility for that purpose was a collateral benefit to the lenders. As to this, Lord Sumption said at [12]:

Lord Sumption went on to explain that the discharge of the existing indebtedness was neither a benefit nor collateral. It was not a benefit to the claimant because the basic measure of damages to which Lord Sumption had referred at [6]:

Mr Salzedo submits to us that the guiding principle was, where a benefit arises in a “but for” sense from the negligence of the defendant, it prima facie comes into account in order to maintain the basic compensatory purpose of damages, subject to rebuttal if the benefit is collateral in the sense that it arose independently of the circumstances giving rise to the loss. He accepted that the benefit must also be legally caused by the wrong, as required by The New Flamenco, but “the only way you can reconcile the three Supreme Court authorities” is to say that a benefit which is factually caused by the wrong is a benefit unless it is collateral or res inter alios acta.

I reject this submission. Although Mr Salzedo seemed to pray Swynson in aid of his submission, it is clear to me that it does not assist him. It was a case about avoided loss, not a benefit for which credit might be required. The claimant’s loan had been repaid and it had therefore suffered no loss. In the present case, AssetCo advanced some £23 million on inter-company loan account to subsidiaries. Those loans were never repaid or discharged for value. They were presumably at some point written off. AssetCo therefore suffered a loss which has not been made good.

Mr Salzedo’s suggestion that the three Supreme Court cases needed to be reconciled suggests some inconsistency between them, although they were decided in quick succession without any adverse comments on the results or reasoning in the other cases. In my view, no reconciliation is required. As Mr Templeman submitted and the judge held, The New Flamenco is clear authority that “but for” causation is insufficient and that, before credit is given for a benefit, it must be shown to have been legally caused by the breach. This was acknowledged by Lord Sumption in Tiuta at [12] where he said that the general rule is that “where the claimant has received some benefit attributable to the events which caused his loss, it must be taken into account in assessing damages” (emphasis added), with the exception of collateral benefits. The reference to benefits being attributable to the events causing the loss encompasses legal, as well as factual, causation. The decision in Tiuta was not concerned with benefits. As Lord Sumption explained at [13], the repayment of the indebtedness under the first facility was not a benefit at all to the claimant. In order to identify the claimant’s loss, it was necessary to look at the whole transaction, the major part of which was to refinance the first facility.

Mr Salzedo’s submission that the only benefits excluded from the quantification of loss are collateral benefits is inconsistent with The New Flamenco, which contains no suggestion that legal causation equates to factual causation in all cases except collateral benefits, and is inconsistent with the passage in Lord Sumption’s judgment in Tiuta at [12] cited above. Indeed, the general principle underlying the exclusion of collateral benefits is that they are benefits “whose receipt arose independently of the circumstances giving rise to the loss” (per Lord Sumption in Swynson at [11]). It is the lack of a causal relationship that underpins their exclusion. Further, I consider the judge was right to say that the exclusion of collateral benefits is directed to the calculation of the claimant’s loss, not to whether credit must be given for a benefit in measuring the damages to be awarded for the claimant’s loss.

I turn therefore to GT’s challenges to the judge’s decision on each of the matters which GT says was a benefit to AssetCo for which credit must be given.

The proceeds of the share issues raise more substantial questions. A preliminary point should be noted. Whether the receipt by a company of the subscription monies for an issue of shares is properly analysed as a benefit for these purposes may not be a straightforward question. AssetCo raised this issue before the judge, submitting that it is not a benefit but a neutral transaction, consistently with (it was submitted) the decision of the High Court of Australia in Pilmer v Duke Group Ltd [2001] HCA 31. The judge noted that, given his findings, the point was academic and, because he did not consider it to have been fully argued before him, he preferred to express no view on it. The point has not been revived before us and the appeal has proceeded on the basis that the receipt of the proceeds of a share issue is a benefit.

Mr Salzedo submits that, if GT is responsible for AssetCo continuing to conduct its business in a dishonest way (or, I would add, on the basis that the business was only “ostensibly sustainable” because of dishonest representations by management which GT should have reported), that too must have been a substantial cause of it raising money to spend in its continuing business. If GT’s wrong caused AssetCo to continue its loss-making activities, the same wrong caused it to raise the money to do so. It was no answer to say, as the judge did, that the proximate cause of the share issue in July 2009 was AssetCo’s need for cash, because its need for cash arose from its continued trading for which GT was, ex hypothesi, responsible.

Mr Templeman submits that it does not follow that, if GT’s breach caused AssetCo’s wasted expenditure on its subsidiaries, it must also have been the cause of whatever benefits it received after the breach. That must depend on the specific circumstances in which the funds were raised. So, if an asset which AssetCo owned before GT’s breach were sold after the breach, the proceeds of sale could not be said to be a benefit caused by GT’s breach. I pause to say that must, of course, be right. The asset could have been sold irrespective of GT’s breach and its sale would have no connection with GT’s breach. Moreover, it is not clear how a sale of an asset at market value could properly be called a benefit for these purposes; by selling the asset, the company would give full value for the cash price it received.

Specifically as regards the share issues, Mr Templeman submits that the effective cause was, as the judge found, that AssetCo needed the money. If one goes further back in the chain of causation, no doubt GT’s breach is the cause of the company continuing to trade on the basis that it was ostensibly sustainable, but while that was the effective cause of the loss, the effective cause of the share issues was the need for cash.

I am unable to accept this distinction. The share issue in July 2009 was undertaken in the course of the continuation of AssetCo’s “ostensibly sustainable” business. It was part and parcel of that business, as Phillips LJ put to Mr Templeman in the course of argument. To say that the need for cash was the effective or proximate cause of the share issue is no different from saying that the effective or proximate cause of AssetCo’s loans to its subsidiaries was their need for cash. In my judgment, the legal cause of both was the audit reports on the 2009 accounts that deprived AssetCo of the opportunity, which it would have taken, of putting a stop to its “ostensibly sustainable” business. The point was well put by GT in its closing submissions, quoted by the judge at [1065], when it said (assuming against itself that its negligence was the legal cause of the losses): “The necessity of raising further funds from equity investors…arose from AssetCo’s financial model which it is common ground would have been known if the audit had been competently performed”.

In my judgment, and accepting that the identification of the legal causation of a benefit, as much as of a loss, involves taking account of all the circumstances to form a common sense overall judgment on the causal nexus between breach and profit or loss (see Famosa Shipping Co Ltd v Armada Bulk Carriers Ltd (The Fanis) [1994] 1 Lloyd’s Rep 633 at 637 per Mance J), the judge’s contrary view displays an error of law and inconsistency in the application of the test of legal causation, leading him to a

The judge also appeared to take account of (what he reasonably took to be) the proximate cause of the investors’ decision to take up the shares, namely the securing of a profitable contract by the group. That was indeed the reason given by AssetCo for the share issue, but the principal purpose was to raise funds to sustain the business by paying urgent debts. Whether the share issue constituted a benefit is determined by reference to AssetCo’s purposes and uses of the funds, not the understandings or motives of the investors in taking up the shares.

This conclusion makes it necessary to consider AssetCo’s submission that considerations of justice, fairness and public policy require that, even if legally caused by GT’s negligence, credit should not be given for the proceeds of the July 2009 share issue. This was the judge’s conclusion, although of course he did not rely on it as he had decided that the share issue was not legally caused by GT’s negligence: see the judgment at [1086]. It is established that such considerations justify the exclusion of collateral benefits, but the effect of the reasoning in Parry v Cleaver on which such exceptions are based will in general be confined to the established categories of such benefits: see Swynson at [11] per Lord Sumption and [100] per Lord Neuberger.

The judge adopted the submissions made to him by AssetCo, which he summarised at [1084] – [1085]. The reasoning was (i) the investors had no claim for their loss against GT (see Caparo), (ii) AssetCo had a claim for the losses caused to it, (iii) if AssetCo had to give credit for the share issue proceeds, it would not to that extent

The premise for this submission is not well-founded. While the investors would have no remedy against GT, they would have a remedy against AssetCo, if (as this argument assumes) they were misled into investing by fraudulent, reckless or negligent statements in, or omissions from, the share offer documents issued by AssetCo. If any such claims had been made, or (perhaps) might realistically have been made, the benefit of the proceeds of the share issue would be matched in whole or in part by a certain or contingent liability and, to that extent, would not be a benefit in the hands of AssetCo. Credit would not therefore, to that extent, be available to GT. In fact, no such claims were made and any such claim was, by the date of the trial, time-barred. In those circumstances, AssetCo received a substantial benefit from its continued trading as a result of management’s dishonest representations to the auditors. I see no reason in principle why GT, whose liability is for negligence, should be denied a credit for that benefit to AssetCo.

For these reasons, I have reached the conclusion that the judge was wrong not to treat the share issue in July 2009 as a benefit to AssetCo, caused as a matter of law as well as fact by GT’s negligence. Credit should therefore be given for AssetCo’s receipt of the proceeds of £7,506,000.

The share issue in March 2011 stands in a rather different position. As the judge noted at [1072] - [1073], the evidence showed that investors did not rely upon the audited accounts but knew that AssetCo and its group were facing serious financial difficulties which the share issue was designed to address. It was a response to the problems caused by AssetCo’s continued operation of an unsustainable business, rather than being part of it. I agree with the judge that this is entirely inconsistent with a sufficient causal nexus between GT’s breaches and this share issue. For the same reason, GT’s argument that the funds raised were applied, in part, in meeting prior liabilities incurred while AssetCo was still operating the business does not assist. That was part of the response to the prior unsustainable trading, not part and parcel of that trading. I should add that, contrary to the judge’s view, I would not regard the passage of time since the 2009 audit as relevant if the true position had not by March 2011 emerged, and in any event AssetCo had a parallel claim in respect of the 2010 audit.

For these reasons, the judge was in my view right not to give credit for the proceeds of the share issue in March 2011.

For the reasons given in this judgment, I would dismiss the appeal, except that the Jaras payment of £1.5 million should not be included in the losses for which GT is liable and the proceeds of the share issue in July 2009, amounting to £7,506,000, is a benefit for which GT is entitled to credit.

I agree.

I also agree.