Primeo Fund (in Official Liquidation) v Bank of Bermuda (Cayman) Ltd and another (Cayman Islands)

This is the judgment of the Board to which every member of the panel has contributed.

This appeal concerns litigation arising from the multi-billion-dollar Ponzi scheme operated by Bernard (‘Bernie’) Madoff (“Mr Madoff”). The appellant, Primeo Fund (in Official Liquidation) (“Primeo”), has brought claims against its former professional service providers, the respondents, for breach of their duties in connection with Mr Madoff’s fraud. This appeal raises a multitude of legal issues, including important questions regarding the law of limitation and the availability of the defence of contributory negligence. The many issues which are addressed in this judgment are set out at paragraphs 50 to 53 below.

Factual background

Primeo is a company (now in official liquidation) which was incorporated in the Cayman Islands on 18 November 1993. It began operations as an open-ended mutual investment fund on 1 January 1994. Primeo raised money from investors, who subscribed for shares, and then invested that money on their behalf. It was promoted, marketed and managed by Bank Austria AG (“Bank Austria”). Primeo was set up as a fund of funds to provide Bank Austria’s customers with access to international investment funds and exposure to the US equity market.

The respondents were professional service providers to Primeo, acting as its administrator and custodian at all relevant times. The respondents were companies within the Bank of Bermuda group, which was acquired by HSBC Bank plc in 2004.

On 21 December 1993, the second respondent, HSBC Securities Services (Luxembourg) SA (“HSBC”), was appointed as Primeo’s custodian and initially as its administrator. The custodian and administrator agreements entered into on 21 December 1993 were superseded on 19 December 1996 by new agreements between Primeo and HSBC (the “1996 Custodian Agreement”), and Primeo and the First Respondent, Bank of Bermuda (Cayman) Ltd (the “Bank of Bermuda” and the “1996 Administration Agreement” respectively) by which the Bank of Bermuda became Primeo’s administrator. Also on 19 December 1996, the Bank of Bermuda and HSBC entered into a delegation agreement, under which the Bank of Bermuda delegated most of its duties under the 1996 Administration Agreement to HSBC. Accordingly, the custodian and administrator function were both, in practice, carried out by HSBC.

Ernst & Young, Cayman (“EY”) was Primeo’s auditor at all relevant times.

Primeo also entered into a subscription agency agreement with Bank Austria. Under an advisory agreement in December 1993, Primeo appointed BA Worldwide Fund Management as its investment adviser, who in turn entered into a sub-advisory agreement with Eurovaleur Inc on 1 January 1994. BA Worldwide Fund Management was replaced as Primeo’s investment adviser by Pioneer Alternative Investment Management Limited, a UniCredit group company, on 25 April 2007. From December 1993, the majority of Primeo’s directors were nominees of the Bank Austria group; and from April 2007, the majority of Primeo’s directors were nominees of the UniCredit group.

In January 1994, and again in February 1996, Primeo entered into agreements with Bernard L Madoff Investment Securities LLC (“BLMIS”) namely Customer Agreements, Trading Authorisations and Options Agreements (together, the “Brokerage Agreements”). Primeo opened two “managed accounts” with BLMIS, which functioned as Primeo’s investment manager, broker and custodian/sub-custodian.

At the outset of its operations, Primeo made a variety of investments, including placing a small proportion of its funds for investment directly with BLMIS (the “Direct BLMIS Investments”). BLMIS purported to operate a “split-strike” investment strategy, which purportedly involved investing in listed securities, treasury bills and cash. Over time, Primeo increased the amount and proportion of its investments with BLMIS. From 2003, Primeo began also to invest indirectly with BLMIS by purchasing shares in other feeder funds associated with Mr Madoff, first Alpha Prime Fund Ltd (“Alpha”), a Bermuda-domiciled investment fund, and later Herald Fund SPC (“Herald”), a Cayman-domiciled fund.

On 1 May 2007, Primeo assigned its rights in respect of its Direct BLMIS Investments to Herald in consideration for new shares in Herald at their reported value (the “Herald Transfer”). Accordingly, from 1 May 2007 onwards, Primeo did not have any direct investments with BLMIS; it was simply a shareholder in Herald and Alpha, which in turn invested with BLMIS.

On 11 December 2008, Mr Madoff surrendered to authorities and was charged with fraudulently operating a multi-billion-dollar Ponzi scheme. Primeo’s board convened the following day and suspended the calculation of Primeo’s Net Asset Value (“NAV”) (a company’s NAV is an expression of the total value of its assets less its liabilities). Primeo was placed into voluntary liquidation on 23 January 2009.

Primeo’s claims

In February 2013, Primeo issued proceedings against the respondents for alleged breach of their contractual duties as Primeo’s administrator and custodian. In summary, in its revised re-re-amended statement of claim, Primeo alleged breaches of duty including the following:

The allegations were denied by the respondents in the Re-Amended Defence, which in brief summary asserted:

A four-month trial commenced in the Grand Court of the Cayman Islands in early November 2016 and oral closing submissions concluded in late February 2017. The court heard evidence from 10 factual witnesses, including former Primeo directors and senior HSBC executives, and 17 expert witnesses, including investment management consultants and professional administrators and custodians. In the trial judgment dated 23 August 2017 (2017 (2) CILR 334), the Honourable Justice Jones QC found that the respondents owed duties to Primeo and breached them. However, he dismissed Primeo’s claims against both respondents on the grounds that (i) the claims infringed the rule against the recovery of reflective loss; (ii) Primeo had not suffered relevant loss for the strict liability claim; (iii) causation had not been proven for the breach of duty claims; and (iv) certain claims were statute barred. The trial judge also found that any award of damages made against the Bank of Bermuda would have been reduced by 75% on account of Primeo’s contributory negligence.

Both Primeo and the respondents appealed various issues to the Cayman Islands Court of Appeal (the “Court of Appeal”). The 10-day appeal hearing took place in late November/early December 2018. In its judgment dated 13 June 2019 (2019 (2) CILR 1), the Court of Appeal allowed Primeo’s appeal in relation to strict liability loss and, in relation to causation, held that the trial judge should have assessed causation in 2005 and 2007 on a loss of a chance basis and overturned his finding that Primeo would have reinvested funds with BLMIS in any event. However, the Court of Appeal dismissed Primeo’s appeal on the basis that its claims were barred by the reflective loss principle. The respondents’ appeal on the issues of duty and breach was dismissed, but HSBC succeeded on its appeal that contributory negligence applied to the claim against HSBC as well as the claim against the Bank of Bermuda. A fuller summary of the findings of both lower courts is set out at paragraphs 25 to 46 below. Before turning to that summary, the following paragraphs provide a brief outline of the claims brought by Herald and Alpha against HSBC.

Herald’s claims

Herald’s administrator and custodian was HSBC. Herald commenced proceedings in Luxembourg against HSBC on 3 April 2009 seeking damages of approximately US$ 2 billion on the basis of: (i) the restitution of securities deposited with HSBC as custodian; (ii) restitution of cash held by HSBC as custodian; (iii) alternatively, a contractual claim for return of the securities; and (iv) an alternative breach of duty claim on the basis of breach of contract, breach of certain statutory duties and an as yet unparticularised tort claim against HSBC as administrator and custodian.

On 22 March 2013, the Luxembourg District Court dismissed Herald’s claim for the restitution of the securities. Both HSBC and Herald appealed aspects of the first instance judgment on Herald’s restitutionary claim. That appeal is ongoing. The remainder of the claim has yet to be determined.

Herald commenced new proceedings in Luxembourg against HSBC and HSBC Bank plc on 29 October 2018 seeking the restitution of approximately US$ 520m plus interest that on Herald’s case (which is denied by HSBC) was transferred by HSBC to BLMIS without transfer instructions from Herald. Apparently due to a translation issue, the summons of 29 October 2018 was served again on 6 November 2018.

On 27 November 2018 (shortly before the 10-year anniversary of Mr Madoff’s arrest, reflecting Luxembourg’s 10-year limitation period), Herald commenced further proceedings in Luxembourg against HSBC. In these proceedings, Herald supplemented its claims for damages, which were the alternative subject of the 2009 summons, including by introducing claims for loss of profits and loss of a chance, and increased the quantum of damages claimed against HSBC to US$ 5.6 billion. These claims have yet to be determined.

Herald did not have any contractual relationship with the Bank of Bermuda and has not issued any claims against the Bank of Bermuda.

Alpha’s claims

Alpha entered into a custodian agreement with Bank of Bermuda Ltd on 12 March 2003. HSBC was Alpha’s sub-custodian. It is common ground between the parties that Alpha did not have any contract with the Bank of Bermuda or HSBC, and that HSBC in fact performed the sub-administration function for Alpha.

Alpha commenced proceedings against HSBC in Luxembourg on 20 October 2009 claiming damages of US$ 346m for breach of contract and in tort. On 28 January 2015, Alpha applied for a temporary suspension of the proceedings, which application was granted by the Luxembourg District Court on 11 February 2015. The proceedings have been suspended since that date.

On 7 December 2018, Alpha commenced further proceedings against HSBC and five other defendants. As against HSBC, Alpha alleges custody and administration failings in terms which closely mirror those made by Primeo in these proceedings. Damages are claimed on various bases, with the principal claim being for US$ 1.16 billion. These claims have yet to be determined and may in due course be consolidated with those commenced by Alpha in 2009.

Alpha has not issued any claims against the Bank of Bermuda.

Summary of findings in the courts below

The following paragraphs summarise the findings of the trial judge and the Court of Appeal in the proceedings below.

Reflective loss

The trial judge found that Primeo’s claims for loss were irrecoverable because they infringed the rule against the recovery of reflective loss. He found that the loss was not separate and distinct from the losses claimed, or capable of being claimed, by Herald and Alpha. He also found that the merits threshold which needed to be met by the claims of Herald and Alpha was “realistic prospect of success” rather than “likely to succeed on the balance of probabilities”, and that those claims had a realistic prospect of success.

The Court of Appeal dismissed Primeo’s appeal on this issue. The Court of Appeal upheld the trial judge’s decision and found that Primeo’s loss was reflective of Herald’s loss and Alpha’s loss, and therefore irrecoverable by Primeo against the respondents.

The custody claim against HSBC: duty

At trial, Primeo argued that, from the entry into the 1996 Custodian Agreement at the latest, HSBC owed duties under that agreement in respect of all of Primeo’s assets, including those invested in BLMIS. HSBC had argued that HSBC never owed duties under the 1996 Custodian Agreement in respect of Primeo’s BLMIS investments and that instead BLMIS owed Primeo custodial duties pursuant to the Brokerage Agreements.

The trial judge found that HSBC owed no safekeeping duty to Primeo in respect of the assets invested in BLMIS before 7 August 2002 because, during that period, BLMIS, rather than HSBC, acted as Primeo’s custodian in respect of those assets under the Brokerage Agreements.

However, the trial judge held that, from 7 August 2002, as part of an “implied tripartite agreement” arising from HSBC and BLMIS entering into the 2002 Sub-Custody Agreement, HSBC owed custodial obligations including a duty to safekeep assets deposited with it, as well as other duties arising under Clause 16(B) of the 1996 Custodian Agreement. Those duties included obligations imposed on HSBC to satisfy itself about the ongoing suitability of BLMIS as its sub-custodian, and to require BLMIS to implement the most effective safeguards available under the laws and practices of the sub-custodian.

The Court of Appeal dismissed the parties’ respective appeals against the position before and from 7 August 2002, and upheld the trial judge’s findings.

The custody claim against HSBC: strict liability loss

The trial judge held that from 7 August 2002 HSBC was, as a matter of construction of the 1996 Custodian Agreement, strictly liable to Primeo for any loss caused by a breach of duty by its sub-custodian BLMIS. However, he found that Primeo suffered no relevant loss for which HSBC was liable, on the basis that BLMIS performed its obligations as sub-custodian by paying cash from time to time in accordance with instructions received from HSBC, and by transferring the total holding standing to the credit of Primeo’s managed account to Herald’s managed account on 1 May 2007, pursuant to the Herald Transfer.

The Court of Appeal allowed Primeo’s appeal on this issue. The Court of Appeal overturned the trial judge’s conclusion on loss, finding that Primeo suffered a loss every time money intended for investment was misappropriated by BLMIS, in breach of the safekeeping duty, and that the Herald Transfer did not extinguish such loss. The Court of Appeal held that quantification of such loss, if the overall outcome of the appeal were that Primeo had established an entitlement to damages on its strict liability claim, would be assessed by a judge of the Grand Court. In its judgment, the Court of Appeal referred to the parties’ submissions regarding the rule in Clayton’s Case (Devaynes v Noble (Clayton’s Case) (1816) 1 Mer. 572, 15 ER 161) and appropriation of payments raised by Primeo in response to the respondents’ argument that Primeo was in any event not entitled to any damages arising from the strict liability claim given that, in the period 7 August 2002 to 2 May 2007, Primeo’s redemptions exceeded its subscriptions by a total of US$25.25m. The Court of Appeal stated they were not in a position to determine the issues arising from those submissions, and that these would need to be referred to the judge assigned to deal with the assessment of damages. The Court of Appeal also held that it would not be right to allow Primeo to seek to enlarge its damages by advancing an additional claim on appeal that HSBC had assumed custodial responsibility for the value of the existing investments and cash held by BLMIS on behalf of Primeo as at 7 August 2002.

The custody claim against HSBC: breach of duty claim

The trial judge found that, from October 2002, HSBC breached its duties under Clause 16(B) of the 1996 Custodian Agreement. The Court of Appeal dismissed the respondents’ appeal and upheld the trial judge’s findings on this issue.

The administration claim against the Bank of Bermuda

Under the terms of the 1996 Administration Agreement, the Bank of Bermuda may only be liable if gross negligence is proved. The trial judge held that the Bank of Bermuda was negligent (but not grossly negligent) in calculating Primeo’s NAV from October 2002 onwards because the Bank of Bermuda, as part of the obligation to calculate an accurate NAV, had failed to verify adequately the existence of assets through the process of reconciliation. He held that a reasonably competent administrator could not have satisfied itself about the existence of assets by reconciling two streams of information received from BLMIS alone and that the Bank of Bermuda’s failure to address the problem of single source reporting was, in the circumstances, negligent. However, the trial judge held that the Bank of Bermuda was grossly negligent from 2 May 2005 onwards. He found that, prior to that date, the Bank of Bermuda had relied upon unqualified audit opinions issued by Primeo’s auditors, EY. However, the trial judge found that, from 2005 onwards, the Bank of Bermuda (through HSBC) knew that EY were no longer willing to rely on the work of BLMIS’ auditor and knew that the custody confirmations issued by HSBC were based entirely on information provided by BLMIS, such that EY’s audit opinion no longer provided any legitimate comfort.

The Court of Appeal dismissed the respondents’ appeal and upheld the trial judge’s findings on these issues.

Causation

The trial judge held that Primeo had not proved on the balance of probabilities that it would have withdrawn its investment with BLMIS if the respondents had complied with their duties, whether in 2002, 2005 or 2007. He considered that Primeo had failed to prove that, but for the custody confirmations, EY would not have continued to issue unqualified audit opinions. He also considered that Primeo’s directors would in any event have continued to invest with BLMIS, for example by investing in other BLMIS feeder funds.

The Court of Appeal upheld the trial judge’s finding on 2002 causation but overruled his findings on 2005 and 2007 causation. As to 2005 and 2007 causation, the Court of Appeal held that the trial judge had, through no fault of his own, applied the wrong legal test when considering whether EY would have issued an unqualified pre-withdrawal audit opinion in the absence of custody confirmations (from HSBC). This was to be assessed on a loss of chance basis by reference to the possible reactions of Mr Madoff and EY in a counter-factual situation rather than, as the trial judge did, on the balance of probabilities.

The Court of Appeal also overruled the trial judge’s finding that, even if Primeo had withdrawn its funds from BLMIS following the lack of a clean pre-withdrawal audit opinion from EY, it would nevertheless have reinvested the money with one or more of the feeder funds associated with Mr Madoff which had not engaged HSBC companies as custodian or administrator. In the Court of Appeal’s judgment, on the balance of probabilities, Primeo would have withdrawn its investment from BLMIS in the absence of an unqualified pre-withdrawal audit opinion and would not have reinvested with one or more of the feeder funds associated with Mr Madoff in such circumstances. Accordingly, the issue would fall for decision by the Grand Court as a matter of quantification of damages by reference to the chances of EY not issuing an unqualified pre-withdrawal audit opinion, thereby causing Primeo to withdraw its funds from BLMIS.

Limitation

The trial judge found that Primeo’s breach of custody and administration duty claims were time-barred before 20 February 2007 (6 years before Primeo issued its claim) but that the custody breach on 23 February 2007 and the administration breaches from February 2007 onwards were not time-barred.

The Court of Appeal upheld the trial judge’s findings on these issues. The Court of Appeal also found that the strict liability claim was not time-barred because BLMIS was to be treated as the agent of HSBC and accordingly its deliberate concealment was to be treated as that of HSBC for limitation purposes.

Contributory negligence

The trial judge held that any damages award against the Bank of Bermuda should be reduced by 75% but that, as matter of law, Primeo’s claim against HSBC was not subject to a defence of contributory negligence.

The Court of Appeal upheld the trial judge’s finding as regards the availability of contributory negligence to the claims against the Bank of Bermuda, and overruled the trial judge’s finding that the defence was not available to HSBC. However, the Court of Appeal also allowed Primeo’s appeal on the amount of the reduction, finding that 50% was the appropriate figure.

Costs

In a costs order dated 9 November 2017, the trial judge ordered Primeo to pay 80% of the respondents’ costs at first instance and to make a payment of US$ 20 million to the respondents on account of such costs.

In the costs ruling dated 15 November 2019, the Court of Appeal ordered Primeo to pay 25% of the respondents’ costs of the appeal and to make an interim payment of US$ 500,000 to the respondents on account of such costs. The Court of Appeal found that it did not have jurisdiction to overturn the trial judge’s first instance costs award unless it allowed the substantive appeal. The Court of Appeal also found that Primeo was unable to seek an adjustment of the trial costs order because the respondents had agreed that Primeo should be relieved from filing a separate appeal notice as to the trial costs order in exchange for Primeo’s agreement that it would only seek to disturb the trial costs order if it succeeded on its substantive appeal in the sense that the appeal was allowed.

Any taxation of costs under both costs orders has been deferred until six months after the determination of Primeo’s appeal to the Board.

Issues to be determined by the Board

Primeo filed its grounds of appeal against the decision of the Court of Appeal on 30 August 2019. Broadly speaking, the issues raised by Primeo’s grounds of appeal are as follows:

The respondents filed their outline of additional grounds for upholding the CICA's decision on 12 November 2019. The following broad issues are raised by the respondents:

The Board gave directions to hear and determine one discrete issue in this appeal, namely the question of whether the rule against the recovery of reflective loss is applicable in principle as a defence to all or part of Primeo’s claim. The hearing of this issue took place on 20 and 21 April 2021. Judgment was handed down on 9 August 2021: [2021] UKPC 22; [2022] 1 All ER (Comm) 1219. The Board humbly advised Her Majesty to allow Primeo’s appeal in relation to the application of the reflective loss rule. The Board’s judgment in Primeo (No 1) is considered in further detail below. In summary, however, the Board (Lord Kitchin and Lord Sales with whom Lord Reed, Lord Hodge and Lord Lloyd-Jones agreed) considered that the rule against the recovery of reflective loss did not bar Primeo’s claims against the Bank of Bermuda and HSBC. The Board’s reasoning was as follows:

The Board has thus settled the reflective loss issue. In this judgment, the remaining issues which arise for determination are addressed in three sections: first, liability and damages issues; second, issues concerning limitation; third, issues relating to contributory negligence.

The following issues are addressed in the liability and damages section of this judgment:

The limitation section of this judgment addresses the following issues:

Finally, the contributory negligence section of this judgment addresses the following issues:

Liability and Damages

The claims against HSBC: did Primeo suffer relevant loss for which HSBC is liable?

HSBC raises a series of points by way of cross-appeal to challenge the ruling by the Court of Appeal that, putting aside the reflective loss defence, Primeo suffered relevant loss for which HSBC is liable. In relation to the strict liability claim against HSBC, the judge held that Primeo suffered no relevant loss for which HSBC was liable (paras 172-177). The Court of Appeal overturned this (paras 182-217). It held that, from August 2002, each time Primeo remitted cash to BLMIS which BLMIS failed to keep safe, in breach of its duties as custodian, Primeo suffered an immediate loss for which HSBC is liable. The reason for this, according to the Court of Appeal, was that each time Primeo remitted cash to BLMIS in relation to which BLMIS purported to hold financial instruments for the benefit of Primeo, BLMIS misapplied the cash so received by using it to prop up the insolvent Ponzi scheme which it was operating. Primeo did not get what it paid for. On each occasion, in return for its payment, Primeo only received rights as against BLMIS of a substantially lower value, ie to participate in its Ponzi scheme and in BLMIS’s insolvency once the scheme was exposed. The Board’s analysis of the reflective loss issue in its judgment in Primeo (No. 1) ([2021] UKPC 22) proceeded on the footing that the Court of Appeal was right about this, but without prejudice to HSBC’s right at the second stage of the hearing of the appeal to challenge the Court of Appeal’s finding that relevant loss had been suffered by Primeo.

Mr Gillis KC, for HSBC, now makes a series of submissions to challenge this aspect of the Court of Appeal’s judgment. The Board will address these in turn below. Before doing so, however, it sets out its view of the matter.

The basic position regarding Primeo’s loss appears to the Board to be straightforward and as identified by the Court of Appeal. Each time Primeo invested cash in BLMIS and BLMIS misappropriated the money, Primeo suffered an immediate and real loss. That loss might be mitigated if Primeo received payments back in respect of the investments it made in BLMIS, but that was a matter of happenstance and does not affect the fact that a real loss was suffered when each cash payment was made to BLMIS and was misappropriated by BLMIS, in breach of its obligation to keep it safe. Primeo accepts that it has to give credit for any payments it received which are attributable to its investments with BLMIS when calculating the damages payable by HSBC in respect of its strict liability claim, but only because such repayments constitute mitigation of the loss which Primeo suffered when it made each investment.

Was there recoverable loss?

Mr Gillis submits that Primeo suffered no recoverable loss. His arguments under this head are variations on the theme that Primeo could not establish causation of loss on a “but for” basis, because even if BLMIS had complied with its safeguarding duty in relation to cash it received in respect of the investments Primeo made directly with it, Primeo would still have suffered the same loss that it in fact suffered. For the proposition of law relied upon, Mr Gillis cited the speech of Lord Nicholls of Birkenhead in Nykredit Mortgage Bank plc v Edward Erdman Group Ltd (No 2) [1997] 1 WLR 1627 (“Nykredit”), at 1631D-E, where Lord Nicholls observed that “but for” causation is assessed by comparing “(a) what the plaintiff’s position would have been if the defendant had fulfilled his duty and (b) the plaintiff’s actual position”.

The Board does not accept Mr Gillis’s general submission nor the arguments grouped under this heading.

Mr Gillis’s basic contention is that it would have been possible for BLMIS to receive all the cash payments made to it by Primeo and keep them safe according to its duty while at the same time operating the same Ponzi scheme. However, in the Board’s view, there was no possibility of this occurring. The only ways in which Primeo’s cash could have been kept safe by BLMIS would have been (a) if the cash (or real financial instruments of equivalent value) had been held by BLMIS on trust for Primeo, thereby avoiding the cash (or the financial instruments) becoming part of the general assets of BLMIS which it was using to carry on the Ponzi scheme and being exposed to loss when BLMIS’s insolvency was discovered, or (b) if BLMIS did not carry on a Ponzi scheme at all.

BLMIS did not in fact hold the cash (or equivalent financial instruments) on trust for Primeo and it was no part of Mr Gillis’s submission that it should have done. That leaves the point that it was implicit in BLMIS’s safekeeping duty that no Ponzi scheme be carried on at all. BLMIS could not, consistently with that duty, both accept Primeo’s cash and carry on the Ponzi scheme, because by doing so it would expose Primeo’s cash (through being part of BLMIS’s general assets) to being lost at any stage when the Ponzi scheme was discovered and collapsed and BLMIS’s insolvency was exposed. The Board accepts the submission of Mr Smith KC for Primeo that BLMIS’s safekeeping duty in relation to Primeo’s cash, in circumstances where the cash was not segregated from that of other investors, encompassed an obligation not to operate a fraudulent Ponzi scheme.

The comparison set out by Lord Nicholls therefore leads to the conclusion that Primeo did indeed suffer loss. Primeo’s position, if BLMIS had complied with its duty, would have been that it suffered no loss in the Ponzi scheme carried on by BLMIS. On the other hand, Primeo’s actual position was that it did suffer the loss of its direct investments in BLMIS through the operation of the Ponzi scheme, since each time BLMIS received the cash invested it misappropriated it for the purposes of that scheme, leaving Primeo with something of far less value, namely the right to participate in BLMIS’s insolvency.

In one variation of his argument, Mr Gillis submitted that even if BLMIS had fulfilled its duty and had continued to hold Primeo’s cash in safe keeping, BLMIS would still have run up the same losses operating its fraudulent Ponzi scheme using other customers’ cash to do so; and still Primeo would only have had an unsecured right to its cash (or equivalent financial securities) and so would have suffered the same loss upon the collapse of the Ponzi scheme as it did in fact suffer. Therefore, he maintained, the breach of duty caused Primeo no loss at all.

In the Board’s view, this submission cannot be sustained. For the reasons given above, the relevant counterfactual “but for” analysis involves asking what would have been the position if BLMIS did not carry on the Ponzi scheme at all. Further, if BLMIS had not fraudulently misled Primeo into believing that its cash was being properly safeguarded and applied in acquiring financial instruments for Primeo’s account, but instead had given a true account of the actual position, Primeo would not have given the cash to BLMIS in the first place or would immediately have required that it be repaid. BLMIS’s duty of safekeeping of Primeo’s cash included not misappropriating it for the Ponzi scheme and not exposing it to loss through including it in the assets at risk from the conduct of the Ponzi scheme. Therefore, to comply with its duty of safekeeping, BLMIS should itself immediately upon receiving cash from Primeo have returned it to Primeo (or, what comes to the same thing, put it in trust for Primeo). If it did not do this, BLMIS should at the very least have given Primeo true information about the position so that Primeo could take steps to demand immediate repayment to prevent its cash from being stolen. It would not have been open to BLMIS, if seeking to comply with the duty it owed Primeo, to keep the cash for itself as an additional fund available to BLMIS’s general creditors upon its insolvency when the Ponzi scheme collapsed, which is the effect of Mr Gillis’s argument.

Furthermore, the Board observes that there is no doubt that Primeo’s cash was in fact misappropriated by BLMIS and used to support the Ponzi scheme. The Board sees no sound reason in principle to make a counterfactual assumption in favour of the fraudster, BLMIS, in this scenario, to the effect that its fraud would have been carried on successfully through to December 2008 in exactly the same way, had it complied with its duty owed to Primeo. This is to invite the Board to lend its authority to give effect to the fraud as perpetrated by BLMIS, rather than to unravel it and give effect to Primeo’s rights according to the true state of affairs as now established on the evidence.

In a further submission, Mr Gillis said that even if BLMIS had kept Primeo’s cash in safekeeping and paid it out on 1 May 2007, immediately prior to the Herald Transfer, Primeo would still have entered the Herald Transfer using that money and so would still have suffered the loss it did in fact suffer; accordingly, for this reason also, it cannot be said that BLMIS’s breach of duty caused loss to Primeo.

The Board cannot accept this argument any more than the previous one. Again, the general point made above applies. It was wholly inconsistent with BLMIS’s safekeeping duty that it should accept Primeo’s cash and at the same time carry on the Ponzi scheme. Alternatively, one might say that on each occasion Primeo made an investment with BLMIS, BLIMIS’s safekeeping duty meant that it had to return the cash (or place it on trust) and/or inform Primeo about the true position and the risk to which it was subject, so that Primeo would have retained the cash. The Ponzi scheme could have been exposed and collapsed at any time. BLMIS could not, consistently with its duty of safekeeping owed to Primeo, have retained the cash it received from investments by Primeo in the period from August 2002, exposing that cash to risk of loss throughout that period, against the possibility of some future repayment to Primeo. Accordingly, the counterfactual position on which Mr Gillis sought to rely for this submission could never have arisen. Moreover, once Primeo appreciated that BLMIS could not keep its cash safe because it was operating a Ponzi scheme, there is no possibility that it would have proceeded to enter into the Herald Transfer and thereby allow its money to continue to be exposed to loss through the collapse of that scheme.

Was breach of BLMIS’s custodial duties rectified by the Herald Transfer?

The judge also held that BLMIS satisfied its custodial duties through the Herald Transfer, whereby Primeo exchanged its choses in action against BLMIS for what it believed were shares in Herald with the same apparent value. The Court of Appeal overruled this part of his judgment as well. Contrary to Mr Gillis’s submission, it was right to do so.

On proper analysis this point is not separate from the previous one. Once it is understood that Primeo suffered a loss each time it made an investment in BLMIS in the period before the Herald Transfer, it is clear that the Herald Transfer did not nullify those losses. The facts that Herald issued shares to Primeo under the Herald Transfer and that as part of that transaction BLMIS made book entries to effect a transfer of the rights of Primeo against BLMIS to Herald in no way put right or rectified BLMIS’s own previous breaches of its safekeeping duty as against Primeo. The purported value of those rights was as fraudulently inflated in Herald’s hands as it had been in Primeo’s hands. On no view could it be said that by participating in the Herald Transfer in this way BLMIS had somehow made safe Primeo’s cash and assets, and thereby belatedly fulfilled its duties as custodian.

Further, Primeo continued to own choses in action against BLMIS, for breach of its safekeeping duty, in respect of the losses Primeo had suffered prior to the Herald Transfer. As the Board observed in its judgment in Primeo (No 1), paras 64-69, Primeo did not waive or abandon its rights against BLMIS in the form of those choses in action when it entered into the Herald Transfer. At paras 209-213 in the present judgment, the Board rejects HSBC’s attempt to advance a new submission that Primeo assigned those choses in action to Herald when it entered into the Herald Transfer.

All that happened under the Herald Transfer was that Primeo exchanged choses in action against BLMIS which had a fraudulently overstated value for shares in Herald which in turn held choses in action against BLMIS which had, in similar fashion, a fraudulently overstated value (thereby causing the value of Herald’s shares to be overstated). Despite the Herald Transfer (indeed, through the mechanism of the Herald Transfer) Primeo remained exposed to the same substantial loss when the Ponzi scheme operated by BLMIS collapsed. The receipt of shares in Herald under the Herald Transfer did not rectify BLMIS’s previous breaches of the safekeeping duty owed to Primeo, nor did it eliminate the losses Primeo had already suffered.

Was there an immediate loss when Primeo’s cash was misappropriated each time it made an investment in BLMIS?

The Court of Appeal held that Primeo suffered immediate and actual loss each time it remitted cash to BLMIS and BLMIS misappropriated the money in order to keep the Ponzi scheme going by making payments against redemptions by other investors with BLMIS. In the Board’s view, the Court of Appeal’s analysis cannot be faulted. On each occasion Primeo made an investment it exchanged cash for worthless (or near worthless) promises by BLMIS, fraudulently giving the appearance of having equivalent worth to the cash invested, and BLMIS misappropriated the cash to prop up the Ponzi scheme it was operating. In these circumstances, in the Board’s view, it is obvious that Primeo suffered real and immediate loss on each occasion it made an investment.

Mr Gillis criticised the Court of Appeal’s analysis by seeking to rely on two authorities, Nykredit and Law Society v Sephton & Co [2006] UKHL 22; [2006] 2 AC 543 (“Sephton”). In his submission, the Court of Appeal erred because, although (as he put it) there is no doubt that an investor such as Primeo suffers a detriment, or possible loss, when investing in a Ponzi scheme, the relevant legal question is when the loss becomes measurable. The investor might be lucky and receive money as full redemption in respect of their investment before the Ponzi scheme collapses. Therefore, according to Mr Gillis, the investor only suffers measurable loss for which damages could be recoverable when the scheme collapses and it becomes clear that they will not receive full redemption.

In the Board’s judgment, this argument is misconceived. On each occasion when Primeo made an investment by paying cash to BLMIS, it clearly suffered loss at that time. It exchanged cash for worthless (or near worthless) assets in the form of promises made by BLMIS, which was insolvent at the relevant time. To test whether or not a loss was suffered by Primeo on each such occasion, the law looks to the underlying reality, not the fraudulent picture presented by BLMIS. There is no difficulty in principle in saying that Primeo suffered real and immediate loss on each such occasion. Nor is there any difficulty in principle in measuring the loss suffered on each such occasion: it was the difference in value between the cash invested by Primeo and the worthless (or near worthless) choses in action as against BLMIS which Primeo received in return (essentially, a right to participate in BLMIS’s insolvency once the true position came to be appreciated, which Primeo was deceived into accepting as consideration for the cash it invested). If Primeo had become aware of the true position on the day it made an investment, it could immediately have sued BLMIS for loss calculated in that way. The fact that BLMIS was successful in deceiving Primeo does not change this legal reality.

Mr Gillis’s reliance on Nykredit and Sephton is misplaced. Mr Gillis submitted, with reference to Lord Nicholls’ speech in Nykredit, that in order to qualify as recoverable loss Primeo’s loss had to be “measurable”, and that Primeo’s loss was not measurable in the relevant sense until the Ponzi scheme collapsed in December 2008. By that time, the Herald Transfer had taken place, Primeo no longer held direct investments in BLMIS but only indirect investments through Herald and Alpha, and so Primeo was barred at that stage by the reflective loss rule from being able to make any claim for damages against BLMIS. But Lord Nicholls’ speech in Nykredit does not support such an analysis.

In Nykredit Lord Nicholls explained the approach to be adopted in a case involving a negligent valuation of security for a lender, in order to determine whether the loss for which the lender claims damages is recoverable and falls within the scope of the duty of care owed by the valuer. The issue in the case concerned the award of interest on damages, which depended on the date when the lender’s cause of action arose in negligence as against the valuer and hence depended on the date when the lender suffered damage for the purpose of that tort. The lender claimed that its cause of action arose when it made the loan (supported by security negligently overvalued by the defendant valuer). The valuer contended that the cause of action arose only after the loan went into default and when the property held as security came to be sold, which was when the lender was visited with the consequence of the valuation being wrong.

Lord Nicholls affirmed (at p 1630D-F) the approach of Stephenson LJ in Forster v Outred & Co [1982] 1 WLR 86, 96-98, in which he accepted the submission of counsel that damage in this context means “any detriment, liability or loss capable of assessment in money terms and it includes liabilities which may arise on a contingency, particularly a contingency over which the plaintiff has no control; things like loss of earning capacity, loss of a chance or bargain, loss of profit, losses incurred from onerous provisions or covenants in leases.” Lord Nicholls then gave as an example (p 1630G) the simple case of someone who buys a house which has been negligently overvalued, who suffers damage at the time of the purchase: “He suffers actual damage by parting with his money and receiving in exchange property worth less than the price he paid.”

In the Board’s view, Lord Nicholls’ statement of the principle and this simple example cover what happened in the present case. With each investment made by Primeo with BLMIS in the relevant period before the Herald Transfer in May 2007 Primeo paid BLMIS cash (which BLMIS immediately misappropriated) and received in exchange property in the form of choses in action (promises by BLMIS to keep the cash safe and to hold a stock of financial instruments of equivalent value for Primeo’s account) which were worth less than the price it paid, a fact concealed by BLMIS by making fraudulent statements about the financial instruments purportedly held by it. Although BLMIS, by its fraud, concealed from Primeo and the world the true state of affairs, the loss suffered by Primeo on each occasion it made an investment was in principle capable of assessment in money terms, by comparing the cash it paid and the value it received in return (ie a share in an insolvent Ponzi scheme, rather than genuine financial instruments safely held to its account).

Given the simplicity of the position in the present case, it is not in fact necessary to focus on the following part of Lord Nicholls’ speech, on which Mr Gillis particularly relied, which deals with the more complicated position which arises in a lending transaction in which the lender takes security which has been negligently overvalued. However, Lord Nicholls’ explanation of the position in that sort of situation, on proper analysis, again supports Primeo’s case on this issue.

The position of the lender in a negligent valuation case such as that in Nykredit is more complicated, as Lord Nicholls explained at p 1631B-C, because the lender looks both to the borrower’s ability to repay under the repayment covenant in the loan agreement and to the security given by the borrower (as valued by the valuer) to ensure that the loan made by the lender is repaid. In one sense the lender suffers a detriment when it makes the loan, with security which has less value than it thinks. But the facts that there are two sources for repayment in such a case and that the security taken may have been intended to include an element of a cushion above the value of the loan means that it is far from certain that the lender will actually suffer any loss at all as events transpire. It is possible that the lender will suffer no loss as a result of the valuer’s negligence if the borrower simply repays the loan and any interest pursuant to the repayment covenant and, moreover, it may be that even when given its true value the security will prove to be sufficient to cover the repayment of the loan. Lord Nicholls therefore addressed the issue in the case in that particular context: “When, then, does the lender first sustain measurable, relevant loss?” (p 1631D). Loss has to be “relevant” in the sense that it is loss which falls within the scope of the duty owed by the valuer (see p 1630F); and it has to be “measurable” in the sense that it can in principle be assessed and given a monetary value (even though that might be subject to the vagaries of what evidence might happen to be available at any given point in time: what Lord Nicholls called “evidential and practical difficulties”, at p 1632B, the existence of which does not prevent a court from finding that loss has been suffered in principle).

Lord Nicholls continued as follows (p 1631D-F):

“The first step in answering this question is to identify the relevant measure of loss. It is axiomatic that in assessing loss caused by the defendant's negligence the basic measure is the comparison between (a) what the plaintiff's position would have been if the defendant had fulfilled his duty of care and (b) the plaintiff's actual position. Frequently, but not always, the plaintiff would not have entered into the relevant transaction had the defendant fulfilled his duty of care and advised the plaintiff, for instance, of the true value of the property. When this is so, a professional negligence claim calls for a comparison between the plaintiff's position had he not entered into the transaction in question and his position under the transaction. That is the basic comparison. Thus, typically in the case of a negligent valuation of an intended loan security, the basic comparison called for is between (a) the amount of money lent by the plaintiff, which he would still have had in the absence of the loan transaction, plus interest at a proper rate, and (b) the value of the rights acquired, namely the borrower’s covenant and the true value of the overvalued property.”

Lord Nicholls then explained that in a valuer case the basic measure of loss might not be the quantum of damages recoverable by the lender, because it can only claim as damages that part of the basic measure of loss which falls within the scope of the valuer’s duty of care, which typically is limited to the amount of the overvaluation: pp 1631H-1632B. This is not a relevant consideration in the present case, as there is no doubt that the loss of Primeo’s cash fell within the scope of the duty owed by BLMIS to safeguard it; so the issue in the present case is whether and when Primeo suffered measurable loss by reason of the breach of that duty, according to the “basic measure”.

Lord Nicholls made it clear that evidential and practical difficulties in assessing the extent of the loss suffered by a lender prior to the security being sold (at which point greater precision becomes possible) do not prevent a court from finding that measurable loss has in principle been suffered at an earlier stage: pp 1632F-1633F. It might be clear that some loss has been suffered, even if it is difficult to assess it with precision. A court would have to do the best it could on the evidence available. As Lord Nicholls pointed out (p 1633C): “In their practical conduct of litigation courts are well able to ensure that assessments of damages are made in a sensible way”.

In giving guidance as to the approach to be adopted in assessing the basic measure of loss, Lord Nicholls said (p 1632C-D):

“Ascribing a value to the borrower's covenant should not be unduly troublesome. A comparable exercise regarding lessees’ covenants is a routine matter when valuing property. Sometimes the comparison will reveal a loss from the inception of the loan transaction. The borrower may be a company with no other assets, its sole business may comprise redeveloping and reselling the property, and for repayment the lender may be looking solely to his security. In such a case, if the property is worth less than the amount of the loan, relevant and measurable loss will be sustained at once. In other cases the borrower's covenant may have value, and until there is default the lender may presently sustain no loss even though the security is worth less than the amount of the loan. Conversely, in some cases there may be no loss even when the borrower defaults. A borrower may default after a while but when he does so, despite the overvaluation, the security may still be adequate.”

In Nykredit itself, on the facts of the case the borrower’s repayment covenant had been shown to be valueless (p 1635A) and, because of the negligent over-valuation, the security was inadequate. Therefore the lender was found to have suffered loss from the date of the loan, as it submitted.

Applying this guidance in the present case, the Board considers that the answer is again clear: with each investment with BLMIS, Primeo’s position if BLMIS had fulfilled its duty would have been that its cash (or genuine financial instruments with the same value) would have been kept safe for its benefit, whereas Primeo’s actual position was that it only obtained a share in an insolvent Ponzi scheme. Looking at the package of rights Primeo received in return for its cash, all it obtained was a chose in action against BLMIS, which was a company which was hopelessly insolvent. Clearly, in the Board’s view, Primeo suffered loss on each occasion that it made such an investment. This was “measurable” loss in the relevant sense, since in principle it is clear that actual loss was suffered even though the quantification of that loss on each occasion would depend on the evidence available when the question came to be addressed at trial and might be lacking in some precision.

Contrary to this, Mr Gillis submits that the issue whether loss was measurable in the present case before the collapse of the Ponzi scheme depends upon the way in which the market would value the promises made by BLMIS to keep Primeo’s cash safe and to make redemption payments upon request. He contends that if Primeo had brought its claim against the respondents in April 2007 (ie shortly before the Herald Transfer) it would only have obtained nominal damages, since Primeo could still at that stage have withdrawn its investments in full from BLMIS, and he says that this highlights why there was no actual loss at the point of investing, only on the collapse of the scheme: only at that time would it be possible to measure the loss suffered by Primeo from investing. The Court of Appeal erred because in undertaking the valuation of the promises by BLMIS at the time when Primeo made its investments the Court of Appeal used hindsight and brought into account the fact that BLMIS was an insolvent Ponzi scheme even though at the time the investments were made that was not known or knowable through normal valuation enquiries. This, said Mr Gillis, was contrary to Lord Nicholls’ approach in Nykredit.

The Board does not accept these submissions. It notes, in the first place, that the effect of them is to invite the Board to give credit to BLMIS for the successful way in which it carried on its fraud, by which it succeeded in concealing the true state of affairs from the world. However, as noted at paras 64 and 73 above, the role of the courts and the Board in this case is to unravel the fraudulent picture presented by BLMIS and to assess whether Primeo suffered loss according to the true underlying facts as at the time of investment, as they are now known to be.

Mr Gillis’s submissions are, in fact, contrary to the approach explained by Lord Nicholls in Nykredit. Two points emerge from his speech. First, when he spoke about valuing the borrower’s covenant (p 1632C-D, quoted above), he was referring to its true value at the relevant time, not such value as might be ascribed to it by the market by reason of a fraudulent concealment of the true state of affairs. He observed that the borrower might be a company with no other assets, so that its covenant was in fact valueless. This involves looking at the true position, not the market perception based on fraudulent statements about the borrower’s financial position. At p 1635A he decided the Nykredit case by stating as a fact that “the borrower’s covenant was worthless”, as determined by looking at events after the loan transaction was entered into (the borrower’s immediate default in meeting the repayment terms). This is in line with Lord Nicholls’ discussion of the simple case of a purchase where a valuer has given negligent advice, in relation to which he made it clear that the existence of loss is to be assessed according to the true state of affairs at the time of the completion of the purchase although the purchaser does not appreciate this and the fact of the loss only comes to light later (p 1630F-G). Although Lord Nicholls (p 1633) and Lord Hoffmann (p 1639) both referred to the potential relevance of the market’s view of a repayment covenant in assessing its value, the Board considers that it is clear that their comments were directed to a situation in which the market is properly informed about relevant matters so that the market valuation of the covenant can be taken to be a good guide as to its true value at the relevant time.

This interpretation of Nykredit is supported by DNB Mortgages v Bullock & Lees [2000] PNLR 427 (CA), in which Robert Walker LJ (as he then was) gave the leading judgment, with which Henry LJ and Scott Baker J agreed. This was another case of a negligent valuation of property provided as security to a lender. An issue of limitation arose and again the question was when the lender suffered damage when making the loan. In between the initiation of negotiations between the lender and the borrower and the making of the loan the borrower had lost his job and become self-employed, but he did not inform the lender of this before the loan transaction was entered into. The change in the position of the borrower, if appreciated, substantially reduced the value of the borrower’s repayment covenant, as the judge found. This had the effect that, applying the guidance in Nykredit and taking the value of the repayment covenant together with the value of the security, the lender suffered loss when it entered into the loan transaction, so that its claim had become time-barred, as the judge held. On appeal the lender submitted that this involved an error of approach, and that the proper approach to valuation of the repayment covenant was to look at how it would have been valued in the general market, which the lender maintained would have been unaware of and unconcerned by the detail of the change in the employment position of the borrower and would have given a higher value to that covenant. The appeal was dismissed. Robert Walker LJ stated (p 436) that the repayment covenant “had to be valued on the evidence available to the deputy judge, restricted though it was, as to the true facts”. This was supported by the decision of the House of Lords in Cartledge v Jopling & Sons [1963] AC 758 (in which a claim in negligence for personal injury was found to be time-barred where the claimant had suffered damage to his lungs before there was any possibility of such damage being detected, and sued only after the existence of the damage became clear) and by the statement of Lord Nicholls in Nykredit, at p 1633D, that the policy of the law “is to advance, rather than retard, the accrual of a cause of action.” The Board agrees with Robert Walker LJ’s analysis.

In the present case, according to the true facts as known at the time of trial, BLMIS was hopelessly insolvent when Primeo made its investments and BLMIS’s promises to keep them safe and to repay were largely valueless. According to Lord Nicholls’ approach in Nykredit, therefore, they are to be treated as such.

The second significant point emphasised by Lord Nicholls in Nykredit is that relevant loss may be identified even though it cannot be precisely quantified on the evidence which is available. See para 82 above and his observations at p 1632B (“The basic comparison gives rise to issues of fact. The moment at which the comparison first reveals a loss will depend on the facts of each case. Such difficulties as there may be are evidential and practical difficulties, not difficulties in principle”) and p 1633B (in a negligent valuation case it is not necessary to wait until the security is sold to be able to tell that a loss has been suffered: “Realisation of the security does not create the lender's loss, nor does it convert a potential loss into an actual loss. Rather, it crystallises the amount of a present loss, which hitherto had been open to be aggravated or diminished by movements in the property market”).

In the present case, as stated above, the Board considers that Primeo clearly suffered loss each time it paid cash to invest with BLMIS and BLMIS misappropriated that cash, even though the extent of BLMIS’s insolvency at the time and the value of its promises to Primeo might have been difficult to assess precisely at each such time. The loss suffered by Primeo was “measurable” in the requisite sense on each such occasion, in that Primeo suffered a loss of economic value which was clear as a matter of principle, and which could have been assessed by a court according to such evidence as might have been available by the time of trial. Accordingly, the Board considers that Mr Gillis’s submission that Primeo only suffered loss when the Ponzi scheme finally collapsed in December 2008 is based on a misunderstanding of the point made by Lord Nicholls regarding the need for “measurable loss”.

For these reasons, the Board considers that Lord Nicholls’ analysis in Nykredit supports Primeo’s case on this point, not that of the respondents. The Board agrees with the submission of Mr Smith for Primeo that in their submissions on this part of the case the respondents confuse the date on which loss is suffered with the date on which the claimant acquires knowledge that it has suffered loss.

The Board turns now to Sephton. This is an authority which, in the Board’s judgment, is completely distinct from the present case. In Sephton, a solicitor had misappropriated clients’ money. He had engaged the defendant firm of accountants to audit returns which he had to make to the Law Society regarding the safekeeping of clients’ funds, and the defendant failed to identify and report on his defalcations. Eventually the solicitor’s misappropriation of clients’ funds was discovered. Clients who were the victims of the fraud made claims for compensation against the Solicitors Compensation Fund operated by the Law Society, and the Law Society paid sums as compensation. The Law Society, in turn, brought a claim against the defendant for damages for negligence, claiming that it had relied on the reports by the defendant to decide not to investigate or intervene in the solicitor’s practice at an earlier stage, which would have enabled it to prevent the misappropriation of client money and hence avoid any claim being made against the Fund. The issue arose whether, for the purposes of limitation, the Law Society suffered relevant loss when it received the defendant’s reports and decided not to act against the solicitor (in which case the claim against the defendant was out of time) or only at the later date when a claim was made against the Fund (in which case the claim was in time). The trial judge held that it was the former, but his judgment was reversed by the Court of Appeal in a decision which was upheld by the House of Lords. The House of Lords held that the misappropriation of funds by the solicitor gave rise to the possibility that a payment might be made by the Fund, contingent upon the loss not otherwise being made good (eg out of the solicitor’s own resources or by teeming and lading, drawing on other stolen client money: see para 9) and a claim being made in proper form against the Fund; however, such a contingent liability, involving the possibility of becoming subject to an obligation in the future to make a payment, did not constitute damage until the contingency occurred and the Fund actually came under an obligation to make a payment to the clients. Accordingly, until a claim was made the Law Society had suffered no loss. The result was that its claim against the defendant had been brought in time.

As Lord Hoffmann pointed out (para 9), the critical question was when damage was suffered by the Law Society. As he explained (para 18), until the contingency of a claim being made occurred the Law Society suffered no relevant damage. All that could be said up to that point was that the defendant’s negligence made it more likely that a claim might be brought against the Fund, but it remained possible that such a claim might never be brought. This was not sufficient to constitute compensable damage in the law of negligence.

By contrast, in the present case Primeo suffered real and immediate loss as soon as it made each investment of cash with BLMIS and the cash was promptly misappropriated by BLMIS. There was nothing contingent about that loss, judged at the time when the investment and misappropriation occurred. Primeo had exchanged hard cash for a share in an insolvent Ponzi scheme. Unlike the position of the solicitor’s clients in Sephton vis-à-vis the Law Society, there was never any doubt that Primeo would want to redeem its investment from BLMIS at some point in time. The fact that BLMIS made a promise to Primeo to hold underlying instruments of equivalent value and to pay money when Primeo wished to redeem its investments meant that, in a relevant sense, Primeo maintained a claim against BLMIS day in, day out to honour its promise to hold such instruments and to be prepared to make redemption payments on request. By contrast, the solicitor’s clients in Sephton asserted no right against the Law Society until they actually made a claim against the Fund. In the present case, unlike in relation to the position of the clients in Sephton, there was no contingency about whether Primeo had a claim against BLMIS on each occasion of misappropriation of its money. The only contingency was whether Primeo’s loss might be made good by repayments which BLMIS might happen to make to it. As a matter of conventional analysis, that is a matter going to whether Primeo’s loss was mitigated and thereby reduced. Unlike in Sephton, it is not a contingency which affects whether Primeo suffered loss in the first place.

Mr Gillis pointed out that if relevant loss was suffered by Primeo each time it invested cash with BLMIS and BLMIS misappropriated that cash, the relevant limitation period for tortious claims would run from that point in time. It would therefore be possible that Primeo could, through the operation of the law on limitation, lose any relevant cause of action it had against BLMIS (and, by reason of BLMIS’s defaults, against HSBC) before it discovered the fraud to which it had been subject. The Board notes that, subject to detailed consideration of the law of limitation and the postponement of any limitation period by reason of BLMIS’s fraud, that might be so. Indeed, Lord Nicholls pointed out in Nykredit at p 1630H that the law of limitation operates in this way, subject to any amelioration by legislation to take account of issues of latent damage (or other matters); Lord Mance likewise made the point in Sephton, at para 56, that a cause of action for damages in tort may accrue without the beneficiary knowing of it; and as noted above Cartledge v Jopling & Sons provides an example of this. These cases emphasise the point that the application of the law of limitation is consequent upon the application of the general law regarding when a cause of action is established, not the other way round. There is no sound argument to postpone the identification of a cause of action according to general law just because the legislation on limitation might have the effect that a claimant is unable to sue in time. In any event, the legislature has made provision to postpone the running of limitation periods in certain circumstances, where fraudulent concealment or mistake is involved. The Board considers the issue of fraudulent concealment by BLMIS and limitation at paras 237-260 below.

Did Primeo suffer its loss in its capacity as a shareholder of Herald on the basis that it was able to withdraw its investments from Herald at all times until December 2008?

This additional ground of appeal by the respondents can be dismissed shortly. It cannot be separated from points (i)-(iii) above. As set out above, and as the Court of Appeal correctly held, Primeo suffered loss each time it made a cash investment with BLMIS in the relevant period of August 2002 to May 2007 and BLMIS misappropriated that money. Clearly, Primeo did not suffer such loss in its capacity as a shareholder in Herald. This additional ground of appeal is an attempt to go behind the Board’s judgment in Primeo (No 1) which cannot be sustained.

Was the loss of cash by Primeo mitigated by sums it received from BLMIS in the period between August 2002 and May 2007?

A fifth submission by Mr Gillis is of an entirely different legal character from those considered above. Even if the Court of Appeal was correct to say that each time Primeo made an investment in BLMIS it suffered an immediate loss, Mr Gillis points to the undisputed fact that in the relevant period between August 2002 and the Herald Transfer in May 2007, during which BLMIS breached the relevant custodial and safekeeping duties in relation to Primeo’s assets, when one takes into account both the cash payments by Primeo to BLMIS and the redemption payments by BLMIS back to Primeo, Primeo suffered no net loss of cash; ie Primeo received back from BLMIS more than it paid to BLMIS. Looking at this period, Mr Gillis submits that each loss suffered by Primeo by paying cash to BLMIS was more than mitigated by payments in the opposite direction. Further, he points out that the only way in which Primeo framed its strict liability claim at trial for damages against HSBC was by reference to the net loss suffered by Primeo in the relevant period. This is true. To meet that submission, Primeo has to be able to say that it was entitled to introduce on appeal to the Court of Appeal a new basis for its damages claim against HSBC arising from its strict liability claim. This contention falls to be assessed under the procedural issues considered below: paras 157-175.

The administration claim against Bank of Bermuda

The Board turns now to the claim by Primeo against the Bank of Bermuda, as administrator, for its failure properly to calculate on each valuation day the NAV upon which Primeo and its shareholders could rely. The trial judge held that the Bank of Bermuda was negligent but not grossly negligent in calculating the NAV from October 2002 because it had failed to take the steps it ought to have taken to verify the existence of the assets purportedly held by BLMIS for Primeo, and that it was grossly negligent from 2 May 2005.

Both parties challenged these findings before the Court of Appeal. Primeo contended that the judge ought to have found that the Bank of Bermuda was grossly negligent from October 2002. The Bank of Bermuda contended that he ought to have found that it was not negligent, let alone grossly negligent, before 2 May 2005; and that it was not grossly negligent after that date.

In the event, the Court of Appeal dismissed all of the challenges to the trial judge’s findings of gross negligence and declined to interfere with his other findings.

On this further appeal, the respondents contend that the Court of Appeal:

The relevant background

As the Board has foreshadowed, Primeo’s claim against the Bank of Bermuda was founded on the relationship between them which was established by the 1996 Administration Agreement. Under the terms of that agreement, the Bank of Bermuda was appointed to act as registrar and accountant of Primeo, and to provide share issue and redemption services under the supervision of the investment adviser. At the same time the Bank of Bermuda entered into a further agreement whereby it delegated the performance of practically all of its obligations as administrator to HSBC. Nevertheless, the Bank of Bermuda accepts that it retained contractual responsibility to Primeo for the performance of these obligations, subject to the qualifications the Board will now explain.

One of the obligations undertaken by the Bank of Bermuda under clause 4.1(xvii) of the 1996 Administration Agreement was to calculate an accurate NAV on which Primeo and its shareholders could rely for the purposes of transacting subscriptions and redemptions, subject to clause 9.2 which relieved it from liability for any act or omission in the course of or in connection with the services it provided in the absence of gross negligence or wilful default on its part or on the part of its servants, agents or delegates.

As the trial judge observed (at para 201), hedge fund administrators are not expected to perform audit procedures, but they are expected to take appropriate steps to satisfy themselves that the published NAV is accurate. The existence of assets such as those the subject of the arrangements giving rise to these proceedings, that is to say, exchange traded securities, is verified by the process of reconciliation; and the pricing of assets is verified by referring to independent pricing services such as Bloomberg. The judge heard unchallenged expert evidence that administrators would normally proceed on the assumption that information received from third party information service providers was reliable, but he held that a reconciliation was still needed for any administrator to satisfy itself about the completeness and the accuracy of the information it had received. As the judge related, there was no criticism of the work done by HSBC to satisfy itself about the reasonableness of the pricing reported by BLMIS. The key issue was whether, in the circumstances arising from the BLMIS business model, a reasonably competent administrator could have satisfied itself about the existence of the relevant assets (apparently credited to Primeo’s managed account) by reconciling two streams of information received from BLMIS alone.

The trial judge heard expert evidence as to the general standards and practices to be expected of a reasonably competent administrator during the relevant period from 1998 to 2008. Mr William Fleury, the Bank of Bermuda’s expert, had direct hands-on experience of the fund administration business at the relevant time. Ms Tanya Beder, Primeo’s expert, had the broader experience one might expect of a consultant. The judge read their expert reports and, as he put it, listened to them give oral evidence at some length. In the end, he preferred the evidence and opinions of Ms Beder. Indeed, the judge formed the view that Mr Fleury’s approach to the issues before him was fundamentally flawed because he was unwilling to recognise that the BLMIS business model presented the Bank of Bermuda with a unique challenge in relation to the verification of assets.

In addressing this issue, the judge explained, at para 217, that in normal circumstances a hedge fund administrator is reassured about the existence of the assets recorded on its clients’ balance sheet by reconciling information provided by two or more independent service providers. Mr Madoff’s approach was very different, however. He insisted that BLMIS must perform the triple function of investment manager, broker and custodian, and BA Worldwide, Primeo’s investment adviser, having apparently accepted that there was a legitimate business reason for this practice, the Bank of Bermuda found itself in the difficult position of having to rely on a single source of information.

Mr Fleury gave evidence to the effect that it was commonplace for administrators to rely on single source reporting. But the trial judge was satisfied that the model Mr Fleury had in mind was very different from that adopted by BLMIS. In the BLMIS model the key functions of investment management, administration and brokerage and custody were purportedly carried on through a single company, wholly owned by one dominant individual who was in a position to override internal controls. Indeed, the entire operation was managed by a group of around 20 people working together in one office. They were responsible for holding the securities in co-mingled accounts, and they performed all the back-office functions. Trade confirmations and month end statements were generated by the same group of people out of the same office, all under the managerial control of Mr Madoff.

This evidence satisfied the trial judge that single source reporting of this kind was unique in the hedge fund industry although it applied not just to Primeo, but also to all the Madoff feeder funds. The judge also found (at para 219) that the relatively high risk of fraud or error inherent in the BLMIS business model must have been manifestly obvious to all concerned, including the fund’s promoters and investment managers and advisers.

Was the Bank of Bermuda negligent from 2002?

In 2002 HSBC, then called Bank of Bermuda (Luxembourg) SA, was represented on Primeo’s Board by Mr Fielding who was Deputy Global head for Client Services for Global Fund Services (“GFS”) for the Bank of Bermuda group of companies with responsibility across all of the jurisdictions in which the group carried on a fund administration and custody business. By 1 October 2002, the GFS Board had informally decided to ask KPMG to undertake audit procedures to provide independent confirmation of the assets held by BLMIS for Primeo and two other clients, referred to as Lagoon and Thema. In the event, KPMG was never engaged. As the judge related, the decision was overtaken by events.

The trial judge recognised that Primeo relied on the failure by the GFS Board to implement its decision to engage KPMG to perform audit procedures in respect of its clients’ balance sheets as constituting the first breach of contract which amounted to wilful default or gross negligence. At this point Primeo was a single manager fund and all of the invested assets were held on the BLMIS managed account.

Nevertheless, the judge was not persuaded that that the failure to take steps to instruct KPMG necessarily meant that the Bank of Bermuda had disregarded or was oblivious to the obvious risks to which the Board was referred, or that it went ahead and issued further NAV calculations conscious of the fact that it was doing so in breach of duty or being reckless in the sense of not caring whether it was acting in breach of duty. He accepted that the Bank of Bermuda was careless in failing to ensure that its decision was not implemented but decided that its carelessness did not amount to gross negligence or wilful default.

The decision having been made that Primeo’s NAV would continue to be prepared on the basis of information supplied by BLMIS without being reconciled with information received from an independent source, the next opportunity to review the matter came in the spring of 2003, and it was discussed at a GFS Board meeting in June, and again at a Board meeting in May 2004. HSBC, as custodian, had not recommended the adoption of any safeguarding features which could provide independent evidence about the existence of the assets held by BLMIS. Nor had the problem of single source reporting been addressed. Nevertheless, BLMIS had continued to perform satisfactorily, and EY had issued another unqualified audit opinion. So, although, as the judge found, the failure to address the single source reporting issue constituted negligence, there had been no material change in circumstances which could justify the conclusion that the Bank of Bermuda was guilty of gross negligence.

On appeal to the Court of Appeal, Primeo submitted that the judge ought to have found that the Bank of Bermuda had been grossly negligent from October 2002. The respondents, on the other hand, contended that he had erred in finding that the Bank of Bermuda had even been guilty of ordinary negligence prior to April 2005. Mr Gillis, for the respondents, submitted, in substance, that the judge was wrong to find that reliance on single source information was something that no competent administrator would do. Everyone knew that the Bank of Bermuda would have to rely on single source information because that was the whole basis of the BLMIS investment model; EY knew that was how the Bank of Bermuda was calculating the NAV and yet they produced unqualified audit opinions in relation to Primeo; and even Ms Beder had not suggested it was negligent for the Bank of Bermuda to rely on single source information prior to 2002.

The Court of Appeal acknowledged the force of these submissions but declined to deal with them, essentially on the basis that it could make no difference to the outcome because, under the terms of the Administration Agreement, the Bank of Bermuda would only be liable if it had been guilty of gross negligence, and this was a matter which the Court of Appeal would consider in assessing its conduct from April 2005.

The respondents now challenge the findings of the judge and the approach adopted by the Court of Appeal. Mr Gillis submits that the judge mischaracterised the nature of the administrator’s role and that his findings were internally inconsistent. He also submits that the Court of Appeal fell into error in not addressing the issue, essentially because the judge’s finding of negligence was central to his approach to the issue of administrator breach generally. He contends that if the respondents are right that the judge’s findings were the result of his failure to understand the administrator’s role then the judge’s later findings of gross negligence necessarily fall away; and further, the Court of Appeal’s approach to the issue of gross negligence wrongly assumed that the respondents were in any event negligent for the reasons the judge had found.

The respondents’ case has at its heart the argument that the judge misunderstood the purpose of ‘reconciliation’ in fund accounting, and that he failed to recognise that it involves no more than the process of identifying discrepancies in the accounts when transactions are checked against holdings, and invoices are matched to payments. This, Mr Gillis continues, is what the Bank of Bermuda was required to do and what it did. Instead, the judge fell into error in accepting Primeo’s argument that reconciliation involves comparing information from two different sources. To suggest that an administrator is required to carry out this kind of reconciliation is a mischaracterisation of its role and wrongly elevates it from one which is essentially administrative into one which involves audit and risk management functions. Mr Gillis also submits that the judge’s finding is inconsistent with his other findings that it was not appropriate to imply an administrator duty independently to verify the existence of Primeo’s assets held by BLMIS; nor was it the role of the administrator to perform managerial and advisory functions. Overall, Mr Gillis continues, the judge’s finding of negligence from 2002 (if not earlier) was in any event plainly wrong and unsupported by the evidence.

The Board is not persuaded by any of these submissions. First, BLMIS presented the Bank of Bermuda with a particular set of problems because it was the only source of the information the administrators needed about transactions and assets. Second, senior individuals within the Bank of Bermuda were by 2002 expressing concern that Mr Madoff presented a significant and increasing risk in the light of BLMIS’s size and the nature of its business. Third, the judge heard evidence from Ms Beder that from 2002, if not earlier, the Bank of Bermuda ought to have acted on the concerns being expressed internally about the lack of independent controls and verification about the information needed to assess the NAV. Fourth, there was a good deal of common ground between the experts that administrators were expected to satisfy themselves that the NAV was accurate and that this required a process of reconciliation. Fifth, for this purpose the administrators needed to be satisfied that the accounting numbers were backed up by assets. Sixth, it was also largely common ground between the experts that the process of reconciliation required the administrators to “get comfortable” that the information was accurate and that the assets really did exist.

The Board is satisfied that in all these circumstances the trial judge was entitled to ask himself what a competent administrator would do when all the information concerning the NAV came from a single source. Until April 2005, the Bank of Bermuda had the benefit of the EY audit opinions, and this gave some comfort as to the existence of the assets underpinning the NAV. However, it did not absolve the Bank of Bermuda from its failure to fulfil its obligations, and this failure, though not grossly negligent in light of the EY opinions, was nevertheless negligent.

The Board is also satisfied that the trial judge was entitled to find the Bank of Bermuda was culpable to this degree until the spring of 2005, and there is no basis upon which it would be appropriate for the Board to interfere with his conclusion. Further, the Court of Appeal made no error in dealing with this aspect of the appeal in the way that it did.

Was the Bank of Bermuda grossly negligent from April or early May 2005?

In April 2005, the position changed. HSBC, in its capacity as custodian, issued to EY custody confirmations of the positions held within the BLMIS managed account as at the year end. EY had requested these custody confirmations earlier in the year after discussions in which they had raised concerns about Mr Madoff, whether the assets attributed to BLMIS really existed and as to the reliability of the United States firm of Friehling & Horowitz (“F&H”) as BLMIS’s auditors.

The trial judge held that now the Bank of Bermuda, as administrator, knew (albeit through HSBC) that EY were no longer willing to rely on the work of F&H and were instead relying on custody confirmations issued by HSBC. But HSBC, in turn, was relying on nothing more than the information issued by BLMIS. The judge held that continuing in this way, knowing that EY were unwilling to rely on F&H’s work and knowing that HSBC had taken no steps to obtain independent confirmation of the existence of the assets, constituted a serious disregard by the Bank of Bermuda, as administrator, of the risks associated with relying solely on information supplied only by BLMIS. This constituted gross negligence.

On appeal to the Court of Appeal, the respondents mounted a sustained challenge to this finding. That challenge foundered because the Court of Appeal had no doubt that the judge had approached the issues before him correctly and that he had made an evaluative assessment of the degree of culpability on the part of the Bank of Bermuda which was unimpeachable. The essential steps in the reasoning of the Court of Appeal were these:

The Court of Appeal also held, in agreement with the judge, that matters did not change after receipt of the first KPMG report in 2006 or the second KPMG report in 2008. Once again, it was understood by all concerned that the work done by KPMG would be and was confined to documents produced by BLMIS. Accordingly, nothing done by KPMG addressed the underlying problem of establishing whether the assets existed, and its work provided no assistance to the Bank of Bermuda on the question to be answered, namely whether it remained grossly negligent to rely only on information derived from BLMIS.

The respondents now contend that the Court of Appeal was wrong to reject their appeal against the judge’s finding that the Bank of Bermuda was guilty of gross negligence from April 2005. Mr Gillis has advanced five main submissions in support of this contention. The Board will address them in turn.

The first is that the Court of Appeal assumed that the Bank of Bermuda had conceded that it was negligent from April 2005 when in truth no such concession had ever been made. The Bank of Bermuda’s primary position was that it was not at any stage negligent but that, even if the judge was entitled to find that it had been negligent from April or May 2005, his finding of gross negligence was plainly wrong and should have been overturned by the Court of Appeal.

The Board does not find this submission persuasive. The Court of Appeal considered with care whether the judge had a proper basis for finding that the Bank of Bermuda was grossly negligent from 2005 and was satisfied that he did. The steps in the reasoning of the Court of Appeal are summarised above and have at their core the proposition that, prior to April 2005, the Bank of Bermuda was able to derive a degree of comfort from the unqualified audit opinions issued by EY. These demonstrated that EY were content that the NAV at the end of the year gave a true and fair view of Primeo’s financial position. But all that changed in April 2005. At that point EY communicated to the Bank of Bermuda that they were no longer content to rely on the work undertaken by F&H and were minded to resign unless they could themselves undertake the appropriate audit work at BLMIS. Thereafter HSBC agreed to provide the relevant custody confirmations and EY relied on them, but the Bank of Bermuda knew that these confirmations were still based entirely on information provided by BLMIS.

The Board is satisfied that this reasoning does not depend on any concession by the Bank of Bermuda that it was negligent (albeit not grossly negligent) from April 2005. It is therefore not necessary for the Board to decide whether that concession was made.

The respondents’ second argument is that the Court of Appeal failed to address their challenge to the judge’s finding of negligence before April 2005. Mr Gillis submits that if the Court of Appeal had considered and accepted the respondents’ argument on their appeal against this finding, then it would (or ought to) have held that the finding of gross negligence after April 2005 also fell away.

The Board does not find this argument any more persuasive than the first and that is so for the reasons the Board has already given, namely that the Court of Appeal made no error in reaching the conclusion that the Bank of Bermuda was grossly negligent from April 2005. Further and in any event, the judge’s approach to the question whether the Bank of Bermuda was negligent from 2002 is unassailable.

The respondents’ third argument is that the Court of Appeal applied the wrong test in assessing the issue of gross negligence, and that the approach it adopted was self-contradictory and wrong. Mr Gillis submits the judge’s finding of gross negligence was based on Mr Fielding’s failure to apply his mind to the question whether HSBC, as custodian, should issue the custody confirmations and that this failure demonstrated he was indifferent to (and so disregarded) the obvious risk of the Bank of Bermuda continuing to prepare the NAV solely on the basis of information derived from BLMIS.

Mr Gillis also submits that the Court of Appeal upheld the decision of the judge on a different basis, namely that Mr Fielding made an assessment of risk that was grossly incompetent, but also found that the judge’s finding of gross negligence could not be faulted. In short, the submission continues, the Court of Appeal failed to grapple with the respondents’ criticisms of the judge’s reasoning or to explain properly any alternative basis for upholding his finding.

The Board does not accept these submissions. There was never any dispute as to the nature of the conduct which might constitute gross negligence. The concept is capable of embracing conduct undertaken with an actual appreciation of the risks involved, but also with a serious disregard of or indifference to an obvious risk, as Mance J explained in Red Sea Tankers [1997] 2 Lloyds Rep 547 at para 208. But of course, whether any particular conduct does or does not amount to gross negligence will depend on all the circumstances and the assessment of the trial judge.

Similarly, the Board does not accept that the judge or the Court of Appeal fell into error in the manner suggested. The judge found (at paras 227 and 228) that Mr Fielding considered that HSBC was entitled to rely on the auditors to do the work necessary to verify the existence of the assets. Whether or not this view was originally misguided, it became untenable once Mr Fielding agreed to provide EY with custody confirmations derived from BLMIS, and from this point, as Primeo contends, Mr Fielding, as the ultimate decision maker, failed to apply his mind to the issue, was indifferent to the obvious risk of continuing to prepare and issue NAVs on the basis of information sourced only from BLMIS, and demonstrated a serious disregard of the risks associated with continued reliance on BLMIS as the only source of the information that EY needed. All of this provided an amply sufficient basis for a finding of gross negligence, and the Board rejects the submission that the Court of Appeal approached the matter incorrectly or arrived at a conclusion which was inadequately explained or materially different from that of the judge.

The respondents’ fourth argument is that the key reasoning of the Court of Appeal is factually flawed. Here Mr Gillis submits that the Court of Appeal noted that the Bank of Bermuda knew that the custody confirmations were based on information derived solely from BLMIS but failed to mention the critical fact that EY also knew the basis on which those confirmations were issued, but were still satisfied they could sign off the accounts without any qualification. Further, Mr Gillis continues, the Court of Appeal overlooked many other factors which the respondents had pointed to as giving comfort as to the reliability of the information derived from BLMIS, and in particular the various matters which the Court had identified when rejecting Primeo’s argument that the Bank of Bermuda was grossly negligent from 2002. In this connection, the respondents also rely on the fact that single source reporting was the inevitable consequence of the BLMIS business model; that BLMIS had always met redemption requests; and that Mr Fielding had conducted a due diligence exercise in 2002 which had raised no specific problems.

The Board does not accept that the Court of Appeal fell into error in the manner for which the respondents contend. The critical point is that the judge found that Mr Fielding and the Bank of Bermuda believed that they could rely on the auditors to do the work necessary to confirm that the assets existed. Further, this was the main factor on which Mr Fielding and, through him, the Bank of Bermuda relied. Nevertheless, once it had been agreed that EY should be provided with the custody confirmations, failing which EY would have had to satisfy themselves that BLMIS did have the assets by auditing BLMIS, Mr Fielding had no possible justification for continuing to rely on EY’s audit opinions as providing any comfort as to the existence of the assets or that BLMIS did indeed have them.

The fifth argument of the respondents is that the Court of Appeal failed to consider whether from April 2005 the Bank of Bermuda could reasonably have exercised its judgment to continue to rely on the information derived from BLMIS.

The Board rejects this argument too, and that is so for the reasons the Board has already explained: the Bank of Bermuda, as administrator, was required to exercise judgment in satisfying itself that the published NAV was accurate, and to do this it needed to take appropriate steps to confirm that the information it had received was complete and accurate, and that the assets did really exist. This was something it failed to do.

Was the Bank of Bermuda grossly negligent from February 2006?

Here the question is whether the Bank of Bermuda remained grossly negligent after receiving the first report from KPMG in February 2006 or the second report from KPMG in 2008.

The judge dealt with this issue in the manner the Board has described. The Court of Appeal was satisfied that he was entitled to find that matters did not change materially after receipt of either of the KPMG reports, and that these reports, whether considered individually or collectively, did not justify the continuation of what was by then a seriously flawed process. It is certainly the case that the reports did not evidence impropriety, but KPMG emphasised that they had relied on the information provided by HSBC and Madoff personnel, and that they had not independently verified it. Further, KPMG had not conducted an audit of HSBC or the information HSBC had provided.

The respondents argue that the judge’s approach to this issue was flawed, as was that of the Court of Appeal, because they both asked the wrong question: they only considered whether the first KPMG report addressed the issues arising from single source reporting. They should have asked whether, after commissioning the first report, the Bank of Bermuda could be said to have conducted a risk assessment that was so deficient as to be a really elementary blunder or to have acted in serious disregard of any obvious risks. What is more, the respondents continue, the judge found that they reasonably took comfort from the first report.

The respondents also contend that the judge and Court of Appeal confused the roles of administrator and custodian and failed to consider whether the Bank of Bermuda could reasonably have exercised its judgment, as administrator, in relying on the information provided by BLMIS, in the light of the first KPMG report.

The Board does not find these submissions persuasive. First, the various matters which led to the conclusion that the Bank of Bermuda was grossly negligent from April 2005 were also relevant at and after the date of the first KPMG report. There had been no material change to them. In particular, as the judge and the Court of Appeal held, none of the matters relied on by the respondents addressed the problem of single source reporting. Further, there was no confusion as to the roles of administrator and custodian. As administrator, the Bank of Bermuda was required to take reasonable care in producing the NAV and to satisfy itself that the NAV was accurate. That in turn required it to take reasonable care to ensure that the assets which formed the basis for that NAV did exist and in circumstances such as those the subject of these proceedings, that required multi-source reconciliation. But that reconciliation was never performed either before the provision of KPMG’s reports or as part of the exercise which KPMG carried out, and the judge was entitled to find that this failure amounted to gross negligence. As Mr Smith submits for Primeo, there was no basis on which a competent administrator could conclude that that the production of the KPMG reports would equate to multi-source reconciliation and that was because all the relevant information came from BLMIS.

Conclusion in relation to the administration claim

For all of these reasons, the Board has reached the firm conclusion that the judge was entitled to find that the Bank of Bermuda, as administrator, was grossly negligent from April 2005 but not before, and that it remained grossly negligent after receipt of the KPMG report in 2006. Further, the Court of Appeal made no error in the way it approached the issues raised by this aspect of the appeal.

Procedural issues

Several arguments, which have been advanced before the Court of Appeal and the Board, have been challenged on the ground that the raising of such arguments in an appellate court is contrary to the principle of finality when the parties have had the opportunity to advance and test the arguments which they chose to advance in a lengthy trial in which the court of first instance has assessed both oral and documentary evidence in some detail. Before addressing the arguments which have been challenged on this ground, the Board discusses the nature of the principle of finality.

There has long been established in the common law a principle that there must be an end to litigation and that a party is not to be vexed by repeated legal challenges in relation to the same subject matter. The court requires parties to bring forward their whole case and, absent special circumstances, will not permit the parties to reopen the same subject of litigation in relation to matters which could have been brought forward in an earlier hearing: Henderson v Henderson (1843) 3 Hare 100, 114-115. The principle has manifested itself in the rules relating to res judicata, cause of action estoppel, issue estoppel and abuse of process. In the present appeal the Board is concerned with the circumstance that a party has sought to raise new arguments on appeal which were not raised before the judge at first instance who presided over the trial.

In Barrow v Bankside Members Agency Ltd [1996] 1 WLR 257 (“Barrow”), 260 Sir Thomas Bingham MR explained the rule in Henderson v Henderson in these terms:

“It requires the parties, when a matter becomes the subject of litigation between them in a court of competent jurisdiction, to bring their whole case before the court so that all aspects of it may be finally decided (subject, of course, to any appeal) once and for all. In the absence of special circumstances, the parties cannot return to the court to advance arguments, claims or defences which they could have brought forward for decision on the first occasion but failed to raise. The rule is not based on the doctrine of res judicata in a narrow sense, nor even on any strict doctrine of issue or cause of action estoppel. It is a rule of public policy based on the desirability, in the general interest as well as that of the parties themselves, that litigation should not drag on for ever and that a defendant should not be oppressed by successive suits when one would do. That is the abuse at which the rule is directed.”

While Henderson v Henderson was not concerned with the matter at issue in this appeal, namely the raising of new points on appeal, the rationale for the rule in that case is relevant also to the raising of new points for the first time in an appellate court.

The adversarial system of justice imposes on the parties the obligation to identify the issues that arise for determination in the litigation so that each party has the opportunity to respond to the points which the other party makes. The function of the judge is to adjudicate on those issues alone: Al-Medenni v Mars UK Ltd [2005] EWCA Civ 1041 (“Al Medenni”), para 21 per Dyson LJ. The lawyers representing each party adduce evidence, both oral and documentary, and cross-examine the witnesses of the other party in order to establish the case which they are advancing and to counter the case which the other party is making. The lawyers in their submissions at the end of the trial address the cases which have been put to the court. In The Owners of the Ship “Tasmania” and the Owners of the Freight v Smith, the Owners of the Ship “City of Corinth” (1890) 15 App. Cas. 223 (“The Tasmania”), 225 Lord Herschell stated:

“The conduct of a cause at trial is governed by, and the questions asked of the witnesses are directed to, the points then suggested. And it is obvious that no care is exercised in the elucidation of facts not material to them.”

As Dyson LJ stated in Al-Medenni, the judge may, in the course of a trial, invite or encourage the parties to recast or modify the issues but must respect a party’s decision if the party refuses to do so. The consequence is that a judge may be compelled to reject a claim on the basis that it was advanced although the judge may think that the claim would have succeeded if it had been advanced on a different basis. In an adversarial system, fairness dictates that outcome. In Air Canada v Secretary of State for Trade [1983] 2 AC 394, 438 Lord Wilberforce stated:

“In a contest purely between one litigant and another … the task of the court is to do, and be seen to be doing, justice between the parties … There is no higher or additional duty to ascertain some independent truth. It often happens, from the imperfection of evidence, or the withholding of it, sometimes by the party in whose favour it would tell if presented, that an adjudication has to be made which is not, and is known not to be, the whole truth of the matter: yet if the decision has been made in accordance with the available evidence and with the law, justice will have been fairly done.”

It is a general rule that a party must advance his whole case at the trial. As Lewison LJ colourfully put it in a case concerning an appeal against a trial judge’s findings of fact: “The trial is not a dress rehearsal. It is the first and last night of the show”: Fage UK Ltd v Chobani UK Ltd, Chobani Inc [2014] EWCA Civ 5; [2014] FSR 29, para 114(ii). There are sound policy reasons for this general rule. First, there is a public interest in the efficient and proportionate resolution of disputes: Sainsbury’s Supermarkets Ltd v Visa Europe Services LLC [2020] UKSC 24; [2020] Bus LR 1196, (“Sainsbury’s”), paras 238-239; and UK Learning Academy Ltd v Secretary of State for Education [2020] EWCA Civ 370 (“UK Learning Academy”), para 44 per David Richards LJ. Secondly, fairness and substantial justice point in the same direction: parties are entitled to know where they stand at the trial and make their decisions relating to the conduct of the litigation in the knowledge of the issues which will be determined at trial: Jones v MBNA International Bank Ltd [2000] EWCA Civ 514 (“Jones”), para 52 per May LJ; parties are not to be vexed by the reformulation of claims in successive suits: Barrow, 260 per Sir Thomas Bingham MR.

These considerations are relevant to new points being taken on appeal. There is no absolute bar on the taking of a new point on appeal. Where the new point is a pure point of law which can be argued on the basis of the facts as found by the judge at first instance, an appellate court may allow the point to be taken if satisfied that the other party has had an opportunity to meet the point and will not suffer prejudice. But an appellate court must exercise great caution before allowing a party to take a new point on appeal after there has been a full trial involving live evidence and cross-examination. In Pitallis v Grant [1989] QB 605 Nourse LJ explained the rule which operates as a norm, quoting from the judgment of Sir George Jessel MR in Ex p Firth, In re Cowburn (1882) 19 Ch D 419, 429:

“the rule is that, if a point was not taken before the tribunal which hears the evidence, and evidence could have been adduced which by any possibility would prevent the point from succeeding, it cannot be taken afterwards. You are bound to take the point in the first instance, so as to enable the other party to give evidence.”

A judge at first instance assesses the parties’ cases and makes findings of fact which are relevant to the cases which the parties have argued in the trial. It has often been said that the trial judge having seen and heard the witnesses is in the best position to assess their evidence on a particular issue in the context of the evidence as a whole. The trial judge also assesses what a document would convey to a reasonable reader in the position of the party who received it having regard to all that preceded it: UK Learning Academy, para 41. The judge’s findings of fact are shaped and limited by the cases which the parties have argued at trial; and there is a danger that injustice may result if an appellate court uses findings which were made in the context of the arguments, which the parties advanced at first instance, to determine a different legal case which a party advances for the first time on appeal.

Against this background, the standard appellate approach, which has been endorsed in later decisions, was set out by Lord Herschell in The Tasmania, 225:

“a Court of Appeal ought only to decide in favour of an appellant on a ground there put forward for the first time, if it be satisfied beyond doubt, first that it has before it all the facts bearing upon the new contention, as completely as would have been the case if the controversy had arisen at the trial; and next, that no satisfactory explanation could have been offered by those whose conduct is impugned if an opportunity for explanation had been afforded them when in the witness box.”

There is, as Snowden J explained in Notting Hill Finance Ltd v Sheikh [2019] EWCA Civ 1337; [2019] 4 WLR 146, paras 26-28, a spectrum of cases. That spectrum ranges from a case where the new point would have caused the parties to conduct the trial differently or would involve a further factual enquiry and prejudice to the opposing party, in which circumstance the principle of finality in litigation carries great weight, to a case where the new point is a pure point of law and the court can be satisfied that the opposing party will not suffer prejudice, in which the appeal court is far more likely to allow the point to be taken. See also Singh v Dass [2019] EWCA Civ 360, paras 16-18 per Haddon-Cave LJ.

The United Kingdom Supreme Court has confirmed this approach in two recent judgments: Sainsbury’s (above) paras 235-243; and Test Claimants in the Franked Investment Income Group Litigation v Revenue and Customs Comrs [2020] UKSC 47; [2022] AC 1 (“FII”), paras 85-94. The cases mentioned above set out important and binding principles regarding what justice requires: AIC Ltd v Federal Airports Authority of Nigeria [2022] UKSC 16; [2022] 1 WLR 3223, para 29, per Lord Briggs and Lord Sales.

In summary, because of the principle of finality in litigation and the policy reasons which support that principle, an appellate court proceeds with great caution before it allows a new point to be taken on appeal. The appellate court may depart from the principle of finality, as it did in Pitallis v Grant, where the point is a pure point of law and the proceedings below would not have been conducted differently if the point had been taken at first instance. It may also depart from the principle in the context of a group litigation where a point of legal principle that will affect many parties is in issue, as in FII (above). But the principle of finality is likely to be upheld if new evidence is needed for its determination or if the opposing party is otherwise prejudiced. The cases which the Board has discussed above provide useful guidance as to the circumstances in which the principle of finality will be upheld.

Against that background the Board considers each of the arguments which have been challenged on the ground that it contravenes the finality principle for an appellate court to consider it. In relation to each argument the Board considers how the parties presented the relevant parts of their cases at first instance before Jones J and in the Court of Appeal and how each of the Cayman courts determined the matters raised by the parties before addressing the challenge based on the finality principle.

Whether HSBC assumed responsibility as custodian for the assets purportedly held by BLMIS on 7 August 2002?

The first argument is raised in Primeo’s appeal to the Board in the context of its claim that HSBC as custodian of Primeo’s investment fund was strictly liable for the acts and omissions of BLMIS acting as HSBC’s sub-custodian under clause 16(B) of the 1993 Custodian Agreement and clause 16(B) of the 1996 Custodian Agreement: see para 32 above. Primeo argued that HSBC was liable for BLMIS’s breach of duty in using the funds it received from Primeo for investment in BLMIS in the operation of the Ponzi scheme. In its case before Jones J, Primeo argued that HSBC had acted as custodian and had incurred strict liability for BLMIS’s breaches of duty through the period from the commencement of the 1994 Brokerage Agreement, when BLMIS opened a management account for Primeo, until 1 May 2007, when the Herald Transfer took effect at which point BLMIS ceased to be a sub-custodian for Primeo. The respondents’ case before Jones J was that HSBC had never been the custodian of Primeo’s investment funds and had therefore not incurred any such liability under those clauses of the 1993 Custodian Agreement or the 1996 Custodian Agreement.

Jones J held (para 157) that it was the common intention of the parties that BLMIS would operate the account in the name of Primeo and would perform the triple function of investment manager, broker and custodian and that this had become an established course of dealing between the parties by January 1996. When Primeo’s capital was restructured into two sub-funds in 1996 and BLMIS thereafter operated two accounts in the name of Primeo, the parties’ intention did not change, “with the result that the parties must have intended that BLMIS would hold the assets credited to both accounts as custodian for Primeo”. This arrangement meant that HSBC had no custodial responsibility for the assets credited to those accounts until the parties altered their contractual arrangement on 7 August 2002 when the 2002 Sub-Custody Agreement was executed. Jones J rejected the respondents’ arguments that this agreement was not valid and enforceable as a matter of Luxembourg law, which governed the agreement. He held (para 159) that the 2002 Sub-Custody Agreement was part of an implied tripartite agreement by which the custody arrangements relating to the managed account assets were restructured. This had the effect of engaging clause 16(B) of the 1996 Custodian Agreement with the result that HSBC remained liable to Primeo for the performance of the custodial duties, whether HSBC or BLMIS as its sub-custodian acted negligently or in wilful breach of duty from 7 August 2002 until the Herald Transfer took place (paras 168 and 171).

This finding by Jones J created a difficulty for Primeo, which had advanced a claim for damages which was measured by the cash which BLMIS had misappropriated since 1993 less all recoveries which Primeo received, giving rise to a net sum which Primeo claimed. In para 172 of his judgment Jones J adopted Primeo’s approach to the quantification of damages and applied it to his finding that HSBC’s liability ran from 7 August 2002 until 1 May 2007. He addressed the question of loss in para 172 of his judgment and stated:

“It is not disputed that BLMIS was conducting a fraudulent Ponzi scheme at all material times and that it wilfully breached its duties as sub-custodian by misappropriating and/or misusing Primeo’s money and covering up its fraud by issuing false statements of account. Mr Smith’s argument is that, for the purposes of strict liability under Clause 16(B), the appropriate analysis of Primeo’s loss is perfectly straightforward. He says that the loss is simply the net amount of cash placed with BLMIS and dissipated in the Ponzi scheme during the relevant period. For these purposes the relevant period runs from 7 August 2002 (when the 2002 Sub-Custody Agreement was executed) until 1 May 2007 (when the Herald Transfer took place and BLMIS’ role as sub-custodian came to an end). The plaintiff’s case is that [HSBC] is liable to make good this loss, subject to giving credit for the actual recoveries which it has received.”

It is clear from Jones J’s judgment that the case which he was addressing was that HSBC was strictly liable for the safeguarding of the funds which it received from Primeo during the period in which it acted as custodian and that in calculating HSBC’s liability, credit had to be given for the recoveries which Primeo received from BLMIS. It is not material to the arguments relating to finality which the Board is addressing that the Court of Appeal disagreed with Jones J when he went on to hold that Primeo suffered no loss as a result of BLMIS’s wilful breach of its duties as sub-custodian (a) because BLMIS performed the repayment instructions received from HSBC and suffered no loss before 1 May 2002 and (b) because when Primeo assigned its rights in the Herald Transfer, it received the full amount of the recorded value of its assets. What is significant to the question of finality is the argument which Primeo advanced before Jones J at first instance in support of its claim arising from HSBC’s role as custodian.

A major difficulty which Primeo faced in the Court of Appeal was that in the period between 7 August 2002 and 2 May 2007 Primeo’s redemptions from BLMIS exceeded its subscriptions by a total of US$25.25 million. On the approach to quantification of damage which Primeo had taken at first instance therefore it suffered no loss during the period in which HSBC was strictly liable for the defaults of BLMIS. Primeo therefore sought to introduce a new case before the Court of Appeal to the effect that on becoming custodian on 7 August 2002 HSBC assumed responsibility to Primeo for the purported value of the funds invested in BLMIS on that date and that HSBC was therefore liable for the differential between the recorded value of the funds and their actual value as a result of their prior misappropriation in the Ponzi scheme.

The Court of Appeal gave this case short shrift. It stated:

“222.

Mr Smith sought to advance not only a claim for damages measured by the cash misappropriated less all recoveries received. He also sought to advance a claim on appeal for the difference between the value of the BLMIS assets for which he said [HSBC] assumed a custodian responsibility when it became Custodian of those assets following the 2002 Sub-Custody Agreement, and their actual value, less any recoveries. At trial, it was Primeo’s primary case that [HSBC] had been custodian of the BLMIS assets with BLMIS as its sub-custodian from the inception of the 1994 Brokerage Agreements and that [HSBC] was liable on a safekeeping basis for the loss in value of the BLMIS assets. However, Mr Smith did not advance below a specific safekeeping damages case tailored to apply to the hypothesis that [HSBC] only became Custodian of the BLMIS assets and BLMIS its sub-custodian from 7 August 2002.

223.

In our judgment, it would not be right to allow Primeo to seek to enlarge its recovery in the manner sought on appeal. Whether Primeo had a safekeeping claim based on custodial responsibility assumed by [HSBC] after 7 August 2002 was a matter to be decided at the trial on such evidence as was relevant to this issue. Such a case not having been advanced below, we hold that Primeo cannot advance it in this appeal.”

Primeo’s case before the Board is that it had advanced the assumption of responsibility argument in the Court of Appeal for the first time in response to the respondents’ case, first advanced in that court, that there was no net loss in the period 7 August 2002 to 2 May 2007 after Jones J had found that HSBC became the custodian only in August 2002. At first instance neither party had pleaded what would have been the net position if HSBC had custodial responsibility for only part of the period from the end of 1993 until 2007 and the parties’ experts had addressed the pleaded cases. In written submissions, Primeo submits that the assumption of responsibility argument is a pure question of law and that any additional evidence can be adduced at a hearing on quantum. Primeo asserts that the parties and the judge at first instance had expressly acknowledged that there would have to be further expert evidence addressing quantum. In his oral submissions, Mr Smith goes further and asserts that further evidence on the assumption of responsibility would not be required.

The Board is not persuaded that it would be just to allow Primeo to advance this new argument on appeal. The Board reaches this view for the following five reasons.

First, the Board is not satisfied that Primeo could not have addressed the quantification of its claim on the hypothesis that the judge might decide that HSBC was custodian for only part of the period between late 1993 and 2007. Indeed, Mr Smith recognised in his closing submissions before Jones J that the judge might decide that HSBC became the custodian only in 2002. Sir Richard Field JA recorded that submission in para 78 of his judgment:

“Primeo submitted in closing that BLMIS was [HSBC]’s sub-custodian continuously from the inception of the 1993 Custodian Agreement down to the Herald Transfer. If this were wrong, BLMIS became [HSBC]’s sub-custodian on the execution of the 1996 Brokerage Agreements; and if this were wrong, BLMIS became a sub-custodian of [HSBC] when the 2002 Sub-Custody Agreement was concluded.” (Emphasis added)

Secondly, Primeo, having foreseen the possibility that the judge might conclude that HSBC was not the custodian before August 2002, should have advanced any arguments about the quantification of its claim and adduced the evidence to support those arguments at the trial. Primeo did not do so but relied on its argument which quantified its loss by reference to all the monies invested in BLMIS under deduction of all monies recovered from BLMIS.

Thirdly, the Board rejects the submission that, in order to mount a case of assumption of responsibility, it would not be necessary to lead further evidence concerning the circumstances in which the 2002 Sub-Custody Agreement and the implied tripartite agreement were entered into and expert evidence on Luxembourg law which governed those agreements.

Fourthly, the Board notes that the attempt to raise the new argument occurred after a trial lasting 49 days which involved live evidence and cross-examination. It is likely that, if the argument about the assumption of responsibility in 2002 for the already misappropriated funds had been raised during the trial, the course of the evidence at the trial would have been different or there would have had to be further factual inquiry. It was incumbent on Primeo to advance the argument during the trial.

Fifthly, the Board is satisfied that the respondents would suffer prejudice if, after such a lengthy trial, the case against HSBC were to be reopened by an examination of the assumption of responsibility argument which Primeo seeks to raise. It would not be just for the Board to allow such a course nor would it be consistent with the efficient resolution of commercial disputes. In short, the principle of finality militates against the presentation of this argument on appeal.

Whether the Court of Appeal erred in directing that questions of appropriation in general, the existence of a running account and whether the rule in Clayton’s case was to apply be referred to the judge assigned to deal with the assessment of damages?

This is the first of the arguments raised in the respondents’ cross-appeal. It arises in the following way. Primeo, having adopted the approach to the quantification of their claim which Jones J set out in para 172 of his judgment (quoted in para 159 above) and faced with the difficulty that it had withdrawn US$25.25 million more from BLMIS than it had invested with it in the period between 7 August 2002 and 2 May 2007, sought to argue before the Court of Appeal (i) that it was entitled to appropriate the sums remitted back to it by HSBC as it saw fit and (ii) that there had been a general running account between it and HSBC and that it could adopt the “first in first out” methodology of Devaynes v Noble (Clayton’s Case). It submitted that it was entitled to allocate the sums which HSBC remitted back to it after 7 August 2002 against credits to the account made before that date.

This argument had not been advanced before Jones J but arose as a result of his finding that HSBC was the custodian of the funds only from 7 August 2002 and the fact that the agreed schedule of payments which was produced at the trial disclosed the net repayments in the period between 7 August 2002 and 2 May 2007. Primeo founded its claim on the broker/custody account numbered IFN092 which, it submitted, operated as a running account. The respondents resisted the claim, submitting that (i) Clayton’s Case was displaced by an implied intention to the contrary, (ii) the rule in Clayton’s Case would not apply because there was not an unbroken or running account, and (iii) the application of the rule in Clayton’s Case would cause injustice because it would render HSBC liable for the pre-2002 losses caused by investment in BLMIS at times when HSBC was not the custodian of the BLMIS assets.

The Court of Appeal, unsurprisingly, stated that it was unable to rule on Primeo’s new submissions and stated (para 227):

“If the overall outcome of this appeal were that Primeo had established an entitlement to damages on its strict liability claim, there would be an order that the damages were to be assessed by a judge sitting in the Financial Services Division of the Grand Court. Since we are not in a position to decide if the broker/custody account was a running account or whether the application of Clayton’s Case would be contrary to the parties’ intentions, or whether in the circumstances of this case Primeo is entitled to appropriate redemption payments as it sees fit, the question as to appropriation in general and whether the rule in Clayton’s Case was to apply would be referred to the judge assigned to deal with the assessment of damages.”

In the Board’s view, the Court of Appeal, in so deciding, failed to give due weight to the principle of finality. Primeo’s closing submissions before the judge had asserted as a fall back that HSBC was strictly liable as custodian from 7 August 2002. The parties had in the “Agreed Issue 2 Schedules”, which were agreed before trial, the information which disclosed the net remittances in the period between 7 August 2002 and 2 May 2007. If Primeo had wished to argue that it had appropriated or could still appropriate the post-August 2002 remittances to its pre-August 2002 investments, it could and should have led evidence on those matters. In Deeley v Lloyds Bank Ltd [1912] AC 756 (“Deeley”), 783 Lord Shaw of Dunfermline quoted with approval the judgment at first instance of Eve J, who stated:

“According to the law of England, the person paying the money has the primary right to say to what account it shall be appointed; the creditor, if the debtor makes no appropriation, has the right to appropriate; and if neither of them exercises the right, then one can look on the matter as a matter of account and see how the creditor has dealt with the payment, in order to ascertain how in fact he did appropriate it”

See also Cory Bros & Co Ltd v Owners of the Turkish Steamship “Mecca” [1897] AC 286 (“Cory Bros”), Lord MacNaghten at p 294. Primeo led no evidence that it had appropriated the remittances to its pre-August 2002 investments nor did it purport to make the appropriation during the trial: see Seymour v Pickett [1905] 1 KB 715, 724, 725 and 727. The rule in Clayton’s Case that the earliest credit is applied to extinguish the earliest debit applies to a running account and is only a presumption of fact which can be rebutted by evidence that the parties had a different intention: Deeley, 771 per Lord Atkinson; Cory Bros, 290 per Lord Halsbury LC and 294 per Lord MacNaghten.

It is clear from those cases that a case based on an appropriation by Primeo or on the application of Clayton’s Case raises questions of fact which were not explored in the trial before Jones J. At trial Primeo advanced a fall-back case that HSBC was liable as custodian from 7 August 2002 but did not lead any evidence to establish its loss on that case other than on a net cash basis. It is not disputed that a case based on appropriation by Primeo would require evidence. Nor is it self-evident that the account upon which Primeo relies is a running account to which the rule in Clayton’s Case would apply unless the parties had a contrary intention. In the Board’s view, Primeo failed to lead evidence at trial in support of its fall-back case in circumstances in which it was clear that the judge might find, as he did, that HSBC did not become a custodian until 7 August 2002. It is significant that the trial before Jones J was a trial of both liability and quantification of damages in which the parties were expected to lead all the evidence in support of their cases on both topics. The Board is not persuaded that it was envisaged that there would be a separate trial on quantification of damages until Jones J made his ruling as to when HSBC was the custodian. Even then, what appears to have been envisaged was that, in the absence of agreement between the experts, there would be a very limited trial to determine whether there was any loss arising from the strict liability claim in the period between August 2002 and May 2007 by reference to the methodology already adopted by the experts. The respondents have asserted that this was the understanding of the parties and Primeo has not contradicted this but has asserted merely that the parties envisaged a trial on quantum.

The Board considers that the Court of Appeal fell into error in failing to give due weight to the principle of finality when it allowed questions of appropriation, the existence of a running account and the application of Clayton’s Case to be considered in a separate trial for the assessment of damages. As it is not disputed that in the period between 7 August 2002 and 2 May 2007 there were net remittances from BLMIS to Primeo, Primeo has failed to establish, on the case which it presented to Jones J, that it suffered a loss arising from HSBC’s liability as custodian in that period.

Whether the Court of Appeal erred in allowing Primeo to advance a new case of a loss of a chance on appeal and whether the judge erred in entertaining the case on auditor causation which Primeo presented in its closing submissions at trial.

At the trial Primeo in its opening submission focused its case on causation on the evidence of Mr O’Neill, who was a director of Bank Austria’s Cayman subsidiary. Primeo’s case was that if the respondents had, in discharge of their contractual duties, informed Primeo’s Board about their concerns regarding HSBC’s ability properly to discharge its obligations as administrator or custodian, the quality of the information which BLMIS provided, or the security or existence of assets held for Primeo, and BLMIS had not addressed HSBC’s concerns in a satisfactory way “then Primeo would have (a) terminated its managed account and withdrawn its investments placed directly with BLMIS, (b) redeemed its shareholdings in Alpha and Herald and (c) would not have invested any further funds with BLMIS”: viz the judgment of Jones J at para 243. Unfortunately for Primeo, Jones J was not prepared to give any weight to the evidence of Mr O’Neill who, he held (para 265), had practically no involvement with Primeo apart from attending three board meetings and who was divorced from Primeo’s decision-making.

In closing submissions, counsel for Primeo, perhaps aware that the judge was not likely to rely on Mr O’Neill’s evidence, advanced cases on causation which had not been foreshadowed in its pleadings or in its evidence. So far as relevant to this appeal the new cases included what has been described as “auditor causation”. The background to this argument is that by 2005 EY, as Primeo’s auditors, were no longer content to rely on the auditing of BLMIS which was conducted by F&H. The argument is that if HSBC had not issued the custody confirmation letter to EY on 5 April 2005, which certified the existence of the assets purportedly being managed by BLMIS, EY would have been forced to resign as Primeo’s auditors because they could not have provided an unqualified audit opinion for the financial year ending 31 December 2004 (and later years) without performing an onsite audit visit at the premises of BLMIS. The argument continues that Mr Madoff would have refused to allow such a visit, as, if he did allow it, it would have revealed the fraud. No other auditor approved by the Cayman Islands Monetary Authority (“CIMA”) would have been prepared to sign off on Primeo’s accounts. Primeo would not have been able to carry on business as a Cayman fund without audited accounts. Primeo would therefore have terminated its relationship with BLMIS.

The judge rejected this case on the balance of probabilities. He was not persuaded (para 261) that Mr Madoff would have refused to give EY access to BLMIS’s books and records, and pointed out that later that year Mr Madoff had allowed KPMG access when it conducted the first of its fraud risk reviews. The judge then addressed the hypothesis that EY were given access to BLMIS’s books and records. He recorded the view that at this point the causation analysis became much more speculative because it turned on the outcome of that hypothesis and there were a range of possible outcomes. He stated (para 262):

“Madoff’s team might have successfully deceived E&Y, as they did in fact deceive KPMG, in which case an unqualified audit opinion would have been issued without the need for any custody confirmation from [HSBC]. In this event, Primeo would have continued placing new money on the managed account. On the other hand, E&Y might have attempted to obtain a confirmation directly from BNY [Bank of New York] which might conceivably have led to the exposure of the Ponzi scheme. In this event, BLMIS would have been put into liquidation and Primeo would have had no opportunity to withdraw its funds or, at least, not without exposing itself to a claw-back claim. Within those two extremes, there is the possibility that E&Y would have come away from BLMIS with an inconclusive result, but without actually suspecting the existence of any fraud or impropriety which would need to be reported to the regulators.”

The judge continued (paras 263 and 264):

“If E&Y (or some other auditor approved by CIMA) had been willing to issue an unqualified audit opinion and continue in office as Primeo’s statutory auditor, the board would not have considered terminating the BLMIS managed account simply because [HSBC] considered the operational risks associated with BLMIS to be unacceptable and were threatening to resign. …

If E&Y had issued an unqualified audit opinion without having any custody confirmation, Primeo would not have withdrawn its investment in BLMIS or redeemed its Herald and Alpha shares.”

The judge therefore rejected Primeo’s case on auditor causation on the balance of probabilities.

HSBC was guilty of further breaches of contract between February and April 2007 when it issued another custody confirmation to EY in circumstances where it was still unable to verify the existence of the managed account assets with information from any independent source. Primeo’s hypothesis of causation was essentially the same as in relation to the breach of contract in early 2005 and the judge held that there was no evidence which made its 2007 causation analysis any less speculative than that applied to 2005. He was not satisfied that Primeo had proved its case on causation.

In its appeal to the Court of Appeal Primeo challenged the judge’s findings on causation. In its skeleton argument Primeo argued for the first time that the matter of auditor causation could be decided either on the balance of probabilities or the assessment of a lost chance. Its primary case was that it was entitled to succeed on the balance of probabilities, but, relying on the judgment of the Court of Appeal of England and Wales in Allied Maples Group Ltd v Simmons & Simmons [1995] 1 WLR 1602, Primeo submitted that, because the auditor causation hypothesis depended upon an assessment of the actions of independent third parties in the counterfactual scenario, it was entitled to succeed in its claim if it could show that there was a real or substantial, rather than a speculative chance that the third parties would have acted in a way which would have avoided the loss which it suffered. In particular, on the hypothesis that HSBC would not have issued custody confirmations in 2005 and 2007, the assessment of whether Mr Madoff would not have given access to BLMIS’s books and papers and whether, with or without such access, EY would not have issued an unqualified audit report on Primeo for the 2004 and 2006 financial years should be made by an assessment in percentage terms of the chance that the third parties would have acted in a way which would have avoided Primeo’s loss. That evaluation of a lost chance would form part of the assessment of the quantum of damage.

The Court of Appeal rejected the argument of the respondents’ counsel that it was not open to Primeo to raise this argument for the first time on appeal. The proper approach to causation in a counter-factual situation was, the Court of Appeal held (para 333), a matter of law which did not need to be pleaded. Because the judge appeared to accept that Primeo would have had no choice but to terminate its managed account with BLMIS if EY did not issue an unqualified audit opinion, the failure of the judge to carry out an assessment of the chances of there being such an unqualified opinion was very significant. The Court of Appeal therefore remitted to the Grand Court to determine in a hearing on the quantification of damages what were the chances of EY not issuing an unqualified audit opinion in 2005 and in 2007 thereby causing Primeo to withdraw its funds from BLMIS.

In a powerful submission to the Board, counsel for the respondents submitted that the judge should not have allowed Primeo to advance the case on auditor causation which had not been pleaded or supported by any written evidence. Counsel also submitted that the Court of Appeal had greatly oversimplified the issues which would have to be addressed in an examination of a lost chance at a quantification trial as there would be a “waterfall” of multiple contingencies to be addressed. This would require the parties to lead significantly different evidence from the evidence at trial. On the evidence led at trial a loss of a chance claim would have failed. Most directly relevant to the question of finality, the respondents argued that the Court of Appeal had not adequately addressed the principle of finality. Primeo had been aware of the loss of chance approach to causation and had advanced it in relation to another part of its claim. Primeo had chosen to raise and argue auditor causation in its closing submissions on a balance of probabilities. Primeo was now seeking “a second bite at the cherry”. Jones J at first instance had decided the issues raised before him but the Court of Appeal, motivated to apply the correct law, had overlooked that a court will not normally decide issues which the parties have not raised. In contrast with Allied Maples, which was decided on a preliminary trial on liability where, depending on the outcome of the liability hearing, a separate hearing on quantum was to be expected, the lengthy Primeo trial had covered both liability and quantification.

The Board is not persuaded by the respondents’ argument based on the written pleadings that the judge erred in law by being willing to address the auditor causation hypothesis at the end of the trial. The Board agrees with counsel for Primeo that, if the respondents had wished to take a pleading point, they should have done so during the trial and thereby given Primeo the opportunity to seek to amend its pleadings should that have been necessary.

Primeo raises a separate argument that it was prejudiced by the respondents’ delay in disclosing some custody confirmations. The Board is not persuaded that this argument takes Primeo anywhere. If it had been Primeo’s intention at the outset of the trial to run an auditor causation case following the disclosure of the custody confirmations and if it considered itself to be prejudiced by a lack of time to assemble evidence for that case, the Board would have expected Primeo to raise the matter with Jones J and, if necessary, seek a postponement of the trial. But it is not evident that Primeo was considering such a case when it sought to recover the custody confirmations and that case did not feature in its opening submissions. For those reasons the Board does not need to address the factual dispute between the parties as to whether Primeo was prejudiced by the production by the respondents shortly before the trial of some custody confirmations when others had been produced earlier.

It is clear from Primeo’s opening submission at trial that Primeo’s primary case at trial was not related to any auditor causation hypothesis but depended on the evidence of Mr O’Neill as to the decisions which the Primeo Board would have taken if informed by the respondents of their concerns about the way in which BLMIS carried on its business. Much of the consideration of the custody confirmations during the trial related to a separate case of estoppel by convention which the judge addressed in paras 165-167 of his judgment and with which the Board is not concerned. The auditor causation hypothesis emerged as a legal argument in Primeo’s closing submissions. It is clear that Primeo was well aware of the loss of a chance approach of Allied Maples where the decisions of third parties have to be assessed in the conduct of a counterfactual assessment of causation. The Board infers that Primeo must have decided to pursue a claim based on the balance of probabilities. This decision may be because, as the respondents assert, Primeo wished to go for “jackpot” damages rather than a reduced sum by reference to the percentage chance of various decisions being taken by various independent third parties. Alternatively, as discussed below, the decision may have resulted from a realisation of the complexity of the assessments to be made and the paucity of the evidence which would support those assessments when the trial had been conducted with a focus on other causal links between the respondents’ breach of contract and Primeo’s claimed loss. It is speculative to attribute a particular motive to the decision to argue the auditor causation case on the balance of probabilities. The important point is that Primeo must have made a conscious decision to present its case in that way.

Had the parties focused on a lost chance approach to auditor causation during the leading of evidence at the trial it is likely that they would have sought to adduce both factual and expert evidence to support or rebut such a case. As the respondents submit, the auditor causation hypothesis involves “a waterfall of contingencies” which include the prospects of BLMIS allowing EY access to its books and papers, the prospect of EY, if granted access (as KPMG was), either discovering fraud or becoming sufficiently suspicious of fraud as not to provide an unqualified audit report, the prospect of EY instead providing a qualified audit opinion, and the prospect of Primeo obtaining a clean audit opinion from another auditor which had been approved by CIMA. These are matters which would each have to be assessed on a lost chance basis, giving rise to different evidence at the trial. It is not known whether witnesses from EY would have been available to give evidence or whether the parties would have obtained documentation from EY Luxembourg. It may be very unlikely that Mr Madoff or one of his associates in BLMIS would have been available to give evidence. But the parties could have adduced expert evidence on relevant auditing practice and standards in order to cast light on the hypothetical behaviour of EY or another auditor on this counterfactual analysis.

The Board considers that if the Grand Court were to address the auditor causation case on a remitted hearing on quantification of damages it would be necessary for the parties to lead evidence on those matters in order for the assessment to be more than mere speculation. There must be a strong possibility that, even with such evidence, the hypothesis would still be too speculative to found a tenable claim. In the Board’s view it would not be fair to give Primeo “a second bite of the cherry” after a lengthy trial on both liability and quantum in which the judge has decided the case on the basis on which it was presented to him. Faced with a case of auditor causation on the balance of probabilities, the respondents could have been confident that they would win that point in view of the number of contingencies which the court had to consider and would not have needed to gather evidence on that matter.

Primeo submits that it was envisaged at trial that there would be a further hearing on quantum. But, so far as the Board can ascertain, that suggestion was confined to a consideration of the experts’ agreed figures of net loss on the strict liability case which the Board has addressed in paras 157 – 169 above if, as occurred, the judge were to conclude that HSBC did not become a custodian until 7 August 2002. A remit of the case to the Grand Court to consider auditor causation on the basis of the loss of a chance is an enquiry of a radically different nature. To allow such an additional trial would be contrary to the public interest in the efficient and proportionate resolution of disputes which the United Kingdom Supreme Court emphasised in Sainsbury’s.

Primeo asserts that the Court of Appeal was correct in holding that the approach adopted in Allied Maples was the correct approach to causation when the actions of third parties are to be assessed in the counterfactual analysis. But that is not the point. Jones J, after a lengthy trial involving both liability and quantification of damages, decided the case on the basis on which the parties presented it to him. If the issue of causation were to be re-opened, there would need to be further evidence as mentioned above and such a further trial would not be an efficient or proportionate way of resolving this dispute. It would unfairly prejudice the respondents and would not be just.

The Board is satisfied that the Court of Appeal erred in failing properly to consider the finality principle and the prejudice which the respondents would suffer if the auditor causation case were to be re-tried. The Board therefore upholds the respondents’ cross-appeal on this ground.

Whether the Judge’s dismissal of Primeo’s causation case for 2005 and 2007 should be upheld for the additional reason that Primeo would have agreed to amend the Custody Agreement.

The respondents argue that the Court of Appeal erred in identifying the wrong counterfactual in relation to the new loss of a chance case.

Jones J held that in August 2002 HSBC was in breach of its contractual duty in clause 16B of the Custodian Agreement to be satisfied as to the ongoing suitability of the sub-custodian and to require the sub-custodian to implement the most effective safeguards to protect Primeo’s assets because it had failed to recommend to Primeo the “available safeguards” which were that BLMIS be required to (i) establish a separate account at the Deposit Trust Company (“DTC”) in which to hold Primeo’s securities and/or to make use of the ID system and (ii) establish a separate account or sub-account with the Bank of New York (“BNY”) in which to hold Primeo’s Treasury bills. (The Board should explain that BLMIS fraudulently represented that DTC held its securities and BNY its Treasury bills). If HSBC had performed its duty and BLMIS had refused to implement the available safeguards, HSBC would have been minded to resign but Primeo, in order to preserve its relationship with BLMIS, would have agreed to introduce a clause into the Custodian Agreement which removed HSBC’s liability for loss arising in relation to assets placed with BLMIS in the absence of fraud, dishonesty, negligence or wilful default on the part of HSBC. The judge held, and the Court of Appeal upheld his judgment, that Primeo had failed on the balance of probabilities to prove that as at August 2002, if HSBC had not been negligent, Primeo would have withdrawn its investment with BLMIS and would not have invested further monies.

By 2005 circumstances had changed as EY had lost confidence in the auditing of BLMIS by F&H. The issue by HSBC on 5 April 2005 of the custody confirmation gave assurance to EY in place of F&H’s audit that the assets purportedly held by BLMIS existed.

The respondents now seek to argue that the Court of Appeal erred in treating as the appropriate counterfactual the hypothesis that HSBC had refused to provide the 2005 custody confirmation. Both the judge (para 255) and the Court of Appeal (para 318) recognised that the issuing of the custody confirmation was not a breach of the Custodian Agreement but was a consequence of HSBC’s further breach of that agreement in not recommending or taking steps towards implementing the available safeguards in 2005. In essence the respondents argue that on the correct counterfactual hypothesis HSBC would have been protected by the amendment of the Custodian Agreement to exclude their liability and would have issued the custody confirmation with the consequence that there would have been no chance of EY not issuing an unqualified audit opinion.

As the Board has rejected the argument that Primeo be allowed at this stage in the litigation to advance a loss of chance case in relation to the auditor causation, it is not necessary to resolve this argument.

Whether the Court of Appeal erred if and to the extent that they made findings on the loss of a chance in 2005 and 2007 in respect of the claim against Bank of Bermuda.

The question which this ground of appeal raises is whether the Court of Appeal’s findings in relation to a loss of a chance case are confined to the case against HSBC or extend to the case against Bank of Bermuda. This depends on the correct interpretation of the Court of Appeal’s judgment. The Board can deal with this ground very briefly.

In the Board’s view there is force in Primeo’s answer to this ground of appeal. The courts below analysed causation in relation to Bank of Bermuda and HSBC together as the breaches were closely linked. The judge found that Bank of Bermuda had been grossly negligent in calculating NAV on the valuation dates starting with 2 May 2005 after HSBC had issued custody confirmations when EY had lost confidence in and were no longer prepared to rely on the auditing of BLMIS by F&H. The audit opinion of EY no longer provided any legitimate comfort to Bank of Bermuda to support its calculation of NAV. The Court of Appeal upheld that finding. The findings on causation were made against the following background. Bank of Bermuda had delegated its duties as administrator to HSBC which carried out both the administration and custodian functions for Primeo. It appears that HSBC recorded the assets purportedly held by BLMIS on software called Geneva in performance of the administration function which was the contractual responsibility of Bank of Bermuda: see the judgment of Jones J at paras 38 and 103. The assets purportedly held by BLMIS were recorded on this software for the purposes of accounting and valuation, and the custody confirmations which HSBC provided in and after 2005 were by reference to a Position Appraisal Report Summary generated by this administration software: see judgment of Jones J para 107. The erroneous Geneva records, which formed the basis of the custody confirmations, would not have existed if Bank of Bermuda and HSBC as its delegate had not negligently relied solely on information provided by BLMIS. The judge (paras 255 and 258) and the Court of Appeal (paras 317-348) dealt with causation in 2005 and 2007 in relation to both Bank of Bermuda and HSBC together.

In the Board’s view the Court of Appeal did not err in considering those causation arguments in relation to Bank of Bermuda and HSBC together. The counterfactual of EY not issuing an unqualified audit opinion is, at least arguably, intimately related to Primeo’s case against Bank of Bermuda as Bank of Bermuda took comfort from the audit opinion in providing the NAVs and it and Primeo might have behaved differently in the absence of such an opinion. But, because the Board has held that the Court of Appeal erred in its allowance of the loss of a chance case contrary to the principle of finality, the argument has no consequence.

Whether the Court of Appeal erred in overturning the judge’s finding that Primeo’s directors would have reinvested monies withdrawn from BLMIS into another Madoff feeder fund after a post-withdrawal audit opinion.

The background to this dispute in relation to causation is the case which Primeo advanced as a counterfactual in its submissions before Jones J on auditor causation. The stages of the argument on causation in 2005, which the Board has discussed above, may conveniently be summarised: (i) the respondents threaten to resign and HSBC does not issue the custody confirmation of 5 April 2005; (ii) EY, dissatisfied with the audit work of F&H and without the comfort of HSBC’s custody confirmation, demands access to BLMIS’s books and papers; (iii) Mr Madoff refuses to give EY such access; (iv) this causes EY to resign as auditor of Primeo and no other auditor approved by CIMA is prepared to accept the engagement as auditor and issue an unqualified audit opinion on Primeo’s financial statements before the deadline of 30 June 2005; (v) as a result, Primeo has to withdraw its investments in BLMIS and thereby avoids the loss of its investments in the Ponzi scheme.

Jones J held (para 259) that it was implicit in Primeo’s argument that EY would have relied on BLMIS’s payment and Primeo’s receipt of the cash balance withdrawn from BLMIS in early 2005, which would be vouched by a bank certificate, as evidence that the assets reflected on Primeo’s balance sheet must have existed as at 31 December 2004. This would have enabled EY to issue an unqualified audit opinion on Primeo’s 2004 accounts.

The judge held on the balance of probabilities that Primeo, which understood the BLMIS business model and its implications, would not have launched an entirely new investment strategy at this stage. Such a course of action would have risked massive redemptions by its investors and consequent loss of fee income. Instead, Primeo would have invested the withdrawn funds in one or more of the other Madoff funds which had not engaged the respondents as custodian or administrator (paras 266-267). The judge applied similar reasoning to the auditor causation case in 2007 (paras 268-276).

The Court of Appeal held that the judge’s conclusion on his evaluation of the counterfactual was outside the reasonable band of decisions open to him and that on a balance of probabilities Primeo would have withdrawn its funds from BLMIS and not reinvested them in some other Madoff feeder fund (para 340).

The respondents challenge this decision, arguing that the Court of Appeal had not demonstrated that it had a sufficient basis for overturning the judge’s evaluation. It is not necessary for the Board to resolve this challenge because it would be relevant only if Primeo were allowed to advance a loss of chance claim in relation to auditor’s causation.

The Board therefore addresses the issue briefly. Primeo challenges the judge’s assertion that it was implicit in its counterfactual that EY issues an unqualified audit report in 2005 after Primeo withdraws its investments from BLMIS. Primeo points out, as it did in its submissions to the Court of Appeal (paras 179-180), that its claims for loss of its investments in the BLMIS Ponzi scheme both before and after 2005 did not depend upon EY having issued such a post-withdrawal audit report. That may be so. But those were not the only claims which Primeo pursued before Jones J: see his judgment, para 243 and the Court of Appeal’s judgment, para 304. Primeo’s claim for loss of future profits on alternative investments did depend upon Primeo being able to invest the funds withdrawn from BLMIS and that would have required Primeo to have had an unqualified audit report. The Board therefore considers that the existence of a post-withdrawal audit report was to that extent implicit in Primeo’s case before the judge. The judge is not to be criticised for reaching the conclusion that such an audit report was implicit in Primeo’s claims. In the event, Primeo did not pursue its claim for future loss of profits before the Court of Appeal.

The focus of the respondents’ challenge is that the Court of Appeal did not consider the materials which were before the judge and which the judge had considered, did not have a nuanced understanding of how Primeo had put its case on causation at trial, and did not hold that the judge had made an error of fact or of law or otherwise adequately explain why it held that the conclusion which he reached on his evaluation was not open to him.

The Court of Appeal gave a succinct summary of its reasons in para 340 of its judgment. It recorded, first, that there was no evidence from a representative of EY or any expert audit evidence of what an auditor would do if Primeo had successfully withdrawn its investments from BLMIS on a date after the 2004 balance sheet date. The Court of Appeal, secondly, held that even if an unqualified audit opinion were obtainable, the respondents as custodian and administrator would be saying that they were concerned that they could not obtain satisfactory evidence that the assets existed, that Mr Madoff had refused to agree to the available safeguards and that the custodian and administrator were going to resign unless their contracts were amended so as to exclude any responsibility on their part for the existence of the assets with BLMIS. The Court of Appeal’s third point was that the auditors were telling Primeo that they were unable to satisfy themselves as to the existence of the assets because Mr Madoff was refusing them access or because having been given access they had not been able to confirm the existence of the assets and would not be able to issue a pre-withdrawal opinion. The court concluded (sub-paragraph (iv)):

“Even allowing for Bank Austria/Primeo’s enthusiasm for Mr Madoff, we find it less than probable that, in these circumstances, having withdrawn the assets from BLMIS, the directors would nevertheless immediately reinvest the funds with BLMIS despite their custodian, their administrator and their auditors all saying that they had been unable to verify that the assets really existed. We consider a contrary conclusion to be outside the band of decisions reasonably open to the trial judge.”

The Board has some difficulty with this reasoning. The second and third points which the Court of Appeal made and the conclusion in para 340(iv) are expressly on the hypothesis not only that Primeo successfully withdraws the funds but also that EY then issues a post-withdrawal audit opinion relying on the withdrawal as evidence of the existence of the assets. On that hypothesis, the auditor’s concerns and those of the custodian and the administrator would have been resolved by Primeo’s successful withdrawal of its funds. It seems to the Board that the Court of Appeal in its concluding sub-paragraph may have been intending to address the counterfactual in which EY are unable to issue an unqualified pre-withdrawal audit opinion and Primeo withdraws their funds but where there is no post-withdrawal unqualified audit opinion.

It is, nevertheless, unnecessary for the Board to resolve this matter. This is because it is not open to Primeo to reargue causation on a loss of a chance approach because it would breach the finality principle. No challenge has been made to the judge’s conclusion that, applying a test of the balance of probabilities, Primeo had failed to establish its auditor causation case. It is important to recall that the judge rejected the third stage of Primeo’s counterfactual analysis, namely that Mr Madoff would have refused access to BLMIS books and papers. He concluded that it was more likely than not that Mr Madoff would have allowed EY access to BLMIS’s books and papers in early 2005 just as he later allowed KPMG such access. In that circumstance, the causation analysis became much more speculative and the range of options which the judge set out in para 262 of his judgment and which the Board quotes in para 178 above would arise.

Whether the Court of Appeal’s dismissal of the claim in relation to the Direct BLMIS Investments (ie Primeo’s rights in respect of its managed account with BLMIS) should be upheld for the additional reason that Primeo assigned to Herald all rights and remedies in relation thereto pursuant to the Herald Transfer.

The respondents raise this argument as an answer to the case, which Primeo has sought to advance in this appeal, that in 2002 HSBC assumed responsibility as custodian for the assets which BLMIS purportedly held as the fruits of Primeo’s investment in BLMIS before 7 August 2002.

The respondents seek to argue that when in April 2007 Primeo assigned to Herald the total holding standing to the credit of Primeo’s managed account with BLMIS in consideration for the issue of shares in Herald, it also assigned to Herald any other claims it had against the respondents. The respondents assert that, as a consequence of that transfer, Primeo is not the correct plaintiff for the losses which it pleads in respect of its direct investments. In oral submissions counsel for the respondents was at pains to emphasise that the point would arise only if the Board were to reject his submissions on the finality of litigation which the respondents plead against Primeo. Mr Gillis did not invite the Board to determine the correct plaintiff issue. He sought to have the matter remitted to the Grand Court for a further hearing only if the Board were not to uphold his submissions on the finality principle. As those submissions have been successful, the point does not arise.

In any event, the submission itself falls foul of the principle of finality. As Mr Gillis recognised, sauce for the goose is sauce for the gander. The respondents did not plead in any way this argument before Jones J, when it could and should have been pursued at trial. If the point had been raised at that stage of the proceedings, the trial judge could have addressed the factual circumstances surrounding the assignation to Herald. This might have assisted in the interpretation of the relevant documents which were Herald’s agreement with Primeo for the issue of the shares in Herald in return for Primeo’s assets in its BLMIS managed account and the instructions, which Primeo gave to BLMIS, to transfer the assets in its BLMIS managed account to Herald’s managed account. It is far from clear whether evidence of the factual matrix of the assignment would have supported an interpretation of that transaction to the effect that, in addition to assigning rights against BLMIS, Primeo assigned its rights against the respondents. But the Board is not invited to determine the point.

If the point had been raised at the trial, this might have involved the adducing of evidence which was relevant to that issue but which otherwise would have been irrelevant to the matters which the parties did raise at the trial. No such evidence was adduced. The point was first mentioned in one sentence in the respondents’ written submission to the Court of Appeal who did not address it in their judgment. It was also mentioned in the respondents’ written case before the Board in the earlier hearing on reflective loss. It was not addressed in the respondents’ outline grounds for this appeal.

It is too late to raise this point now.

Limitation

The writ in this action was issued on 20 February 2013. The respondents have asserted a limitation defence in respect of any causes of action which arose more than six years earlier, that is, prior to 20 February 2007. In response, counsel for Primeo rely upon section 37 of the Limitation Act, which postpones the commencement of the ordinary six-year limitation period for contractual claims in specified circumstances. It provides, so far as material:

“(1)

Subject to subsection (3), where in the case of any action for which a period of limitation is prescribed by this Act, either –

(a)

the action is based upon the fraud of the defendant;

(b)

any fact relevant to the plaintiff’s right of action has been deliberately concealed from him by the defendant; or

(c)

the action is for relief from the consequences of a mistake;

the period of limitation does not begin to run until the plaintiff has discovered, or could with reasonable diligence have discovered, the fraud, concealment or mistake. References in this subsection to the defendant include references to the defendant’s agent, and to any person through whom the defendant claims, and his agent.

(2)

For the purposes of subsection (1), deliberate commission of a breach of duty in circumstances in which it is unlikely to be discovered for some time amounts to deliberate concealment of the facts involved in that breach of duty.”

Subsection (3) is concerned with the protection of bona fide purchasers for value without notice.

The courts below held that:

On this appeal, Primeo challenges the conclusion of the Court of Appeal as to (1). In its cross-appeal, HSBC challenges the Court of Appeal’s conclusion as to (2). We shall consider each issue in turn. Since section 37 of the Limitation Act is identical in all material respects to section 32 of the English Limitation Act 1980, the Board will treat the English authorities on the construction of the latter provision as relevant also to the construction of the Cayman provision.

Whether reckless breaches of contract amount to “deliberate commission of a breach of duty” within the meaning of section 37(2) of the Limitation Act.

In relation to this issue, Primeo’s argument can be summarised as follows:

In reply, the respondents argue:

The ordinary meaning of “deliberate”

As a matter of ordinary language, there is no doubt that the adjectives “deliberate” and “reckless” have different meanings. The Board respectfully disagrees with the contrary view expressed in Canada Square by Rose LJ (para 94). For example, the Concise Oxford Dictionary defines “deliberate” as meaning “done consciously and intentionally”, and reckless as meaning “without thought or care for the consequences of an action”. Those definitions capture the distinction between the two words in ordinary speech.

In legal contexts, recklessness can have different shades of meaning, as has often been noted: see, for example, O (A Child) v Rhodes [2015] UKSC32; [2016] AC 219, paras 84, 87 and 113. But as far as the Board is aware, and as far as counsel’s researches have disclosed, it has never been treated as a synonym of “deliberate”. The point is illustrated by the case of Grant v International Insurance Co of Hanover Ltd [2021] UKSC 12; [2021] 1 WLR 2465, where the United Kingdom Supreme Court had to decide whether an insurance exclusion for “liability arising out of deliberate acts” by employees applied to reckless acts. The court held that it did not. Lord Hamblen, giving a judgment with which the other members of the court agreed, stated at para 52:

“First, the starting point is the natural meaning of ‘deliberate’ acts. This connotes consciously performing an act intending its consequences. It involves a different state of mind to recklessness.”

For similar reasons, in O v Rhodes the United Kingdom Supreme Court held that the tort of intentionally causing physical or psychological harm does not extend to recklessness. As Lord Neuberger stated at para 113, “[i]ntentionality … excludes not merely negligently harmful statements, but also recklessly harmful statements”.

In legislation, “reckless” is often employed in conjunction with “deliberate”, or other words signifying knowledge or intentionality, in order to widen the ambit of the provision. Examples drawn from Cayman legislation include section 3(b)(ii) of the Workmen’s Compensation Act (1996 Revision) (“committed deliberately or with a reckless disregard”), sections 18(5) and 23(1)(b) of the Terrorism Act (2018 Revision) (“knowingly or recklessly”), and section 3(3)(b) of the Advance Passenger Information Act (2018 Revision) (“intentionally or recklessly”).

The meaning of “deliberate” in the context of section 32 of the 1980 Act was discussed by the House of Lords in Cave v Robinson Jarvis & Rolf [2002] UKHL 18; [2003] 1 AC 384 (“Cave”). The case concerned a different issue from the present case, but the judgments nevertheless contain guidance as to how the issue in the present case should be resolved. The question was whether the commission of an intentional act which was negligent constituted the deliberate commission of a breach of duty within the meaning of section 32(2) of the 1980 Act, which, as we have explained, is in substantially the same terms as section 37(2) of the Limitation Act. In other words, did the adjective “deliberate”, in the phrase “deliberate commission of a breach of duty” in section 32(2), describe the commission of the act or omission which resulted in the breach of duty, or the commission of the breach of duty itself? The House of Lords held that that the latter was the correct interpretation.

Lord Millett, with whom Lord Mackay of Clashfern and Lord Hobhouse of Woodborough agreed, stated at para 17:

“The question is whether the words ‘deliberate commission of a breach of duty’ in section 32(2) of the 1980 Act mean ‘deliberate commission of [an act or omission, being an act or omission which gives rise to] a breach of duty’, or simply mean ‘deliberate breach of duty’. If the latter, then they refer only to a breach of duty which has been committed intentionally.” (emphasis added)

Lord Millett favoured the latter interpretation (paras 24-26), with the consequence that the words which we have emphasised applied.

Lord Scott of Foscote, with whom Lord Slynn of Hadley, Lord Mackay and Lord Hobhouse agreed, stated at para 58 that the words “deliberate commission of a breach of duty” in section 32(2) of the 1980 Act “exclude a breach of duty that the actor was not aware he was committing”. He added at para 60:

“If the claimant can show that the defendant knew he was committing a breach of duty, or intended to commit the breach of duty – I can discern no difference between the two formulations; each would constitute, in my opinion, a deliberate commission of the breach – then, if the circumstances are such that the claimant is unlikely to discover for some time that the breach of duty has been committed, the facts involved in the breach are taken to have been deliberately concealed for subsection (1)(b) purposes”.

The last two sentences of para 61 are to the same effect, explaining that “the clear words of section 32(2)” show that Parliament has made a distinction “between the case where the actor knows he is committing a breach of duty and the case where he does not”.

Accordingly, the authority of Cave is sufficient by itself to justify the rejection of the argument that recklessness as to whether a breach of duty has been committed is sufficient for section 32(2) of the 1980 Act or section 37(2) of the Limitation Act to apply. For the sake of completeness, however, the Board will also address the remaining arguments.

A restatement of the old law?

Primeo’s argument is essentially that section 32 of the 1980 Act was intended to be a restatement, in modern language, of the old law of concealed fraud, set out in section 26 of the Limitation Act 1939 (2 & 3 Geo 6, c 21) (“the 1939 Act”) and the case law on that provision. The same is true, mutatis mutandis, of section 37 of the Limitation Act. Under the old law, counsel for Primeo argue, a reckless breach of duty was sufficient to prevent the limitation period from running. The difficulty confronting this argument is that the House of Lords twice held that section 32 is not to be construed as if it re-enacted section 26 of the 1939 Act, or the old law of concealed fraud: on the contrary, the House of Lords made it clear that the law was changed by section 7 of the Limitation Amendment Act 1980, which was consolidated in section 32 of the 1980 Act.

The first of the cases in the House of Lords was Sheldon v RHM Outhwaite (Underwriting Agencies) Ltd [1996] AC 102 ("Sheldon”), where the question was whether, and if so how, section 32(1)(b) applied where the deliberate concealment of a relevant fact occurred after the claimant’s cause of action had accrued. That is not a point which arises in the present case, but it is the guidance given by the House to the approach to the construction of section 32 which is in point. The defendants argued that section 32 was the statutory successor of section 26(b) of the 1939 Act, which, it was said, was a statutory enactment of the equitable doctrine of concealed fraud, under which the subsequent concealment of facts did not constitute concealment by fraud. Therefore, it was argued, section 32 should be construed so as to have the same effect. The argument was rejected.

Lord Keith of Kinkel began his analysis at p 140 with the observation that the 1980 Act was a consolidation statute; and “[r]ecourse to the antecedents of a consolidation statute should only be had when there is a real difficulty or ambiguity incapable of being resolved by classical methods of construction: Farrell v Alexander [1977] AC 59, 73, per Lord Wilberforce”. Lord Browne-Wilkinson, with whom Lord Keith agreed, also rejected the argument. He noted at p 144 that the 1980 Act was a consolidating Act, and said that “unless there is an ambiguity, it is not permissible to construe consolidating Acts in the light of their statutory history.” He added (ibid) that “much of the difficulty in this case is raised by the investigation of the statutory history and the decisions of the courts on earlier statutes”: an observation which equally applies to the present case. He went on at p 145 to explain that, even if it were legitimate to look at the legislative history, the immediate predecessor of section 32 of the 1980 Act was not section 26 of the 1939 Act but section 7 of the Limitation Amendment Act 1980, which had deliberately departed from the old law of concealed fraud.

That reasoning was followed in Cave. After describing the effect of section 26(b) of the 1939 Act, Lord Millett stated at para 19:

“Section 32(1)(b) and section 32(2) of the 1980 Act were designed to clarify and, if necessary, change the law by removing all reference to fraud and substituting the more appropriate concept of ‘deliberate concealment’. In such circumstances reference to the antecedent statute and case law is of limited value, since there can be no assumption that the later statute merely reproduced the pre-existing law.”

Lord Scott said at para 46 that the importance of the Sheldon case was “that it insists that if the language of section 32 is clear, effect must be given to that language without regard to the section’s legislative history”. Following that approach, he stated at para 58:

“I would start by adopting the approach prescribed by Lord Browne-Wilkinson in the Sheldon case. Unless there is some ambiguity in the statutory language, recourse to legislative history is unnecessary and impermissible. The relevant words in section 32(2) are ‘deliberate commission of a breach of duty … amounts to deliberate concealment of the facts involved in that breach of duty’. These are clear words of English.”

Counsel for Primeo point out that the speeches in Cave did not address the question whether “deliberate” included recklessness, and that Lord Millett cited at para 20, without adverse comment, a passage from the judgment of Lord Denning MR in King v Victor Parsons & Co [1973] 1 WLR 29, 33-34, in the course of which the Master of the Rolls noted that “fraud”, in the context of section 26(b) of the 1939 Act, had been held to include recklessly committing a wrong as well as doing so knowingly. After citing another case concerned with section 26(b) of the 1939 Act, Lord Millett said at para 23:

“As I have explained, in enacting the 1980 Act Parliament substituted ‘deliberate concealment’ for ‘concealed fraud’. This is a different and more appropriate concept. It cannot be assumed that the law remained the same.”

Accordingly, the change effected in 1980 was not merely the adoption of more modern language: there was a conceptual change in the law. That observation answers counsel’s argument.

In short, it is clear from Sheldon and Cave both that it is impermissible in principle to rely on the 1939 Act and the cases applying it, where the meaning of “deliberate” in the current legislation is clear, and also that the current legislation was intended to have a different effect from the 1939 Act, making reliance on the old case law in order to construe that word inappropriate even if it were permissible.

Policy considerations

In the circumstances, it is unnecessary to address the competing arguments in relation to policy considerations in detail. Put shortly, counsel for Primeo relied on Lord Millett’s observation in Cave, para 7, that “a plaintiff ought not to find that his action is statute-barred before he has had a reasonable opportunity to bring it”. However, the decision of the House of Lords in that case itself demonstrates that that observation is not (and, in its context, was not intended to be) a summary of the law. The case concerned an allegation of professional negligence on the part of a solicitor who had acted for the plaintiff nine years earlier. The House of Lords decided that a negligent breach of duty, in circumstances in which it was unlikely to be discovered for some time, did not fall within the scope of section 32(2) of the 1980 Act and therefore did not postpone the running of the limitation period – notwithstanding that, as a consequence, a plaintiff might find that his action was statute-barred before he had had a reasonable opportunity to bring it. Lord Millett explained at para 15 why that result was considered to be justified, notwithstanding its consequences for the plaintiff. Referring to an earlier case in which the Court of Appeal had reached the contrary conclusion, he said:

“The effect of Brocklesby v Armitage & Guest [(Note) [2002] 1 WLR 598] is to deprive a professional man, charged with having given negligent advice and who denies that his advice was wrong let alone negligent, of any effective limitation defence. However stale the claim, he must defend the action on the merits … This subverts the whole purpose of the Limitation Acts. The harshness of the rule is evident. In the absence of any intentional wrongdoing on his part, it is neither just nor consistent with the policy of the Limitation Acts to expose a professional man to a claim for negligence long after he has retired from practice and has ceased to be covered by indemnity insurance.”

Similar observations as to the justification for confining section 32(2) to cases where the defendant was aware of his own deliberate wrongdoing were made at paras 24 and 27.

These considerations are also relevant where the defendant’s breach of duty is reckless rather than negligent. The point can be illustrated by considering the decision of the House of Lords in R v G [2003] UKHL 50; [2004] 1 AC 1034, para 41,where it was held, in the context of construing section 1 of the Criminal Damage Act 1971, that “[a] person acts recklessly …with respect to … a result when he is aware of a risk that it will occur; and it is, in the circumstances known to him, unreasonable to take the risk”. Many professional people knowingly take risks. For example, a surgeon operating on a gravely ill patient is likely to be well aware of a risk that he may cause the patient’s death. A lawyer advising on a difficult point of law will almost certainly be aware of a risk that his advice may be wrong, and may result in his client’s suffering economic loss. If the R v G test of recklessness were to apply in the context of section 32(2) of the 1980 Act, or section 37 of the Limitation Act, these professional people would have no protection against claims for the indefinite future, whenever it was argued that it had been unreasonable for the surgeon, or the lawyer, to take the risk. Effectively, they would be deprived of the protection of the limitation periods laid down in the 1980 Act and in the Limitation Act where they had arguably made an error of judgment. The implications for professional practice would be drastic, with indefinite exposure to stale claims, long after indemnity insurance had expired.

Canada Square

It follows that Canada Square was wrongly decided on this point. The United Kingdom Supreme Court has recently heard an appeal against the decision of the Court of Appeal in that case, and its judgment on that appeal considers fully the reasoning of the Court of Appeal, and explains why it was mistaken: [2023] UKSC 41.

Conclusions

For all the foregoing reasons, and particularly in the light of the authorities which we have cited, Primeo’s argument must be rejected. Given that conclusion, it is unnecessary to consider the extensive submissions which the Board heard in relation to the pre-1980 law or to consider whether it is permissible to have regard to the Parliamentary material on which reliance was placed (matters discussed in the Supreme Court’s judgment in the case of Canada Square). Nor is it necessary to consider a number of subsidiary arguments which would have arisen if Primeo had succeeded in their submissions on this point. It follows that Primeo cannot rely on section 37(2) of the Limitation Act, so any fault-based causes of action which arose prior to 20 February 2007 are time-barred.

Whether BLMIS is to be treated as HSBC’s agent in respect of the strict liability claim against HSBC, so that its deliberate concealment of relevant facts can be attributed to HSBC for the purposes of section 37(1)(a) and (b) of the Limitation Act

Introduction

Primeo’s strict liability claim against HSBC was brought after the ordinary limitation period had expired in respect of a significant part of it. In light of the findings of the courts below, that claim covers the period from the inception of the Sub-Custody Agreement on 7 August 2002, whereas the ordinary limitation period under the Limitation Act goes back only to 20 February 2007. In order to meet HSBC’s defence of limitation, Primeo relies on section 37(1)(a) and (b) of the Limitation Act, which were set out at para 214 above. Section 37(1)(a) applies where “the action is based upon the fraud of the defendant”. Section 37(1)(b) applies where “any fact relevant to the plaintiff’s right of action has been deliberately concealed from him by the defendant”. Primeo’s action is not based upon the fraud of HSBC, but upon the fraud of BLMIS. Primeo does not suggest that any relevant fact was deliberately concealed from it by HSBC, but that there was such concealment by BLMIS. It is common ground that there was fraud and deliberate concealment of relevant facts by BLMIS. In these circumstances, Primeo’s response to HSBC’s limitation defence relies upon the final sentence of section 37(1):

“References in this subsection to the defendant include references to the defendant’s agent, and to any person through whom the defendant claims, and his agent.”

The critical question is therefore whether BLMIS was HSBC’s “agent”, within the meaning of section 37(1).

Primeo advanced an argument based on the “agency” extension in section 37(1) before the judge, but he did not deal with it. Primeo advanced the argument again in the Court of Appeal, which accepted it in so far as it was based on the deliberate concealment provision in section 37(1)(b): paras 453-458. The Court of Appeal accordingly held that the limitation period did not begin to run until Primeo could with reasonable diligence have discovered BLMIS’s concealment, which meant that no part of Primeo’s strict liability claim was time-barred.

The Board should observe at the outset that although it is accepted that there was deliberate concealment of relevant facts by BLMIS, the Court of Appeal’s approach at para 444 to the meaning of the words “deliberately concealed”, based as it was upon the reasoning of the English Court of Appeal in AIC Ltd v ITS Testing Services (UK) Ltd (The Kriti Palm) [2006] EWCA Civ 1601; [2007] 1 All ER (Comm) 667, should be read subject to the disapproval of that reasoning by the United Kingdom Supreme Court in Canada Square. That point does not, however, in itself affect the Court of Appeal’s conclusions, or its reasoning in relation to the meaning of “agent”.

Mr Gillis, for HSBC, submits that the Court of Appeal erred because, based on English Court of Appeal authorities in relation to the predecessor provision in section 26(b) of the (UK) Limitation Act 1939, it interpreted the word “agent” in section 37(1) to include any independent contractor engaged by the primary obligor. In Mr Gillis’s submission, the word “agent” should be given its conventional legal meaning. Mr Smith’s riposte for Primeo is that, accepting that the word “agent” should be interpreted according to its conventional meaning in general law, BLMIS acted as HSBC’s agent as so understood.

The meaning of “agent” in the limitation statutes

In agreement with the High Court of Australia, the Board would accept as a general rule that terms with a technical legal meaning, when used in legislation, can be taken to have been used by the legislature in their legal and technical sense, unless a contrary intention appears: Bathurst City Council v PWC Properties Pty Ltd (1998) 195 CLR 566; [1998] HCA 59, para 45; applied to the term “agent” in Guan v Lui [2021] NSWCA 65, para 42.

The formula used in section 37(1) is in line with that used in section 32(1) of the (UK) Limitation Act 1980, which derives from section 26(a) and (b) of the (UK) Limitation Act 1939. In all three provisions, the word “agent” is not defined, which indicates that the legislature intended it should bear its usual meaning according to the general law.

Section 26(a) and (b) of the 1939 Act were enacted to give effect to the recommendations of the Law Revision Committee’s Fifth Interim Report on Statutes of Limitation (1936) (Cmd 5334). So far as material, section 26 postponed the commencement of the limitation period where “(a) the action is based upon the fraud of the defendant or his agent or of any person through whom he claims or his agent, or (b) the right of action is concealed by the fraud of any such person as aforesaid”.

The Law Revision Committee had observed (para 22) that although, as a general rule, it is no answer to a defence of limitation to say that the claimant was unaware of the existence of the cause of action until after the expiration of the statutory period, it would be unjust for a defendant to rely upon a lapse of time created by his own misconduct where either the cause of action springs from the fraud of the defendant or its existence has been concealed by its fraudulent conduct. The approach to limitation in courts of equity where there had been fraudulent concealment by the defendant should be adopted as the general approach, including in relation to claims at law. The Committee recommended that where a cause of action is founded on fraud “committed by the defendant or his agent” or where a cause of action unconnected with fraud is fraudulently concealed from the claimant “by the defendant or his agent” time should not begin to run for limitation purposes until the claimant first discovered such fraud or could with reasonable diligence have discovered it.

The Committee said nothing to suggest that the word “agent” here should have any meaning other than the conventional legal meaning. Previously in its report the Committee used the term “agent” with that meaning: see its discussion at para 19 of the fresh accrual of a right of action upon acknowledgement of a debt.

In the same way, the 1939 Act used the term “agent” in its legal sense in section 24, which concerned the fresh accrual of rights of action on an acknowledgment or part payment under section 23, and, following the Committee’s recommendation, provided by subsection (2):

“Any such acknowledgment or payment as aforesaid may be made by the agent of the person by whom it is required to be made under the last foregoing section, and shall be made to the person, or to an agent of the person, whose title or claim is being acknowledged or, as the case may be, in respect of whose claim the payment is being made.”

In the context of that provision, the agent has the capacity, by making an acknowledgment or part payment, to affect the legal position of the person by whom it is required to be made; or, by receiving the acknowledgment or part payment, to affect the legal position of the person whose title or claim is being acknowledged, or in respect of whose claim the payment is being made. “Agent” is being used in its ordinary legal sense, as an agent is, as explained below, a person with the capacity to alter his principal’s legal relations with third parties. There is no indication that the same word was intended to bear a different meaning in section 26.

Section 32 of the (UK) 1980 Act and section 37(1) of the (Cayman) Limitation Act follow and adopt the same scheme as section 26(a) and (b) of the 1939 Act. The meaning of “agent” in both of them is the same.

The background to the 1980 Act, on which the Cayman Limitation Act is based, was the review of the law of limitation by the Law Reform Committee in its 21^st Report (Final Report on Limitation of Actions), Cmnd 6923 (1977). One of the mischiefs identified by the committee in that report was the disparity between the language of section 26 of the 1939 Act and the way in which it had been construed by the courts, as more fully explained by the Supreme Court in Canada Square: [2023] UKSC 41at paras 50-51.

The 1980 Act uses the word “agent” in two provisions. The first is section 30(2), which corresponds to section 24 of the 1939 Act (para 246 above) and to section 34 of the (Cayman) Limitation Act. In that provision, the word clearly refers to an agent in the ordinary sense of the term, as explained at para 246 above. The second provision is section 32(1), where the final sentence is in the same terms as the final sentence of section 37(1) of the Limitation Act. In section 32(1), as in section 30(2), the 1980 Act provides for the conduct of the defendant’s agent to affect the legal relationship between the defendant and a third party, namely the plaintiff. That is a context in which one would expect the word “agent” to bear its conventional legal meaning: see para 250 below. That is also the meaning which one would expect it to have in its application to section 32(1)(a), which applies where “the action is based upon the fraud of the defendant”. In addition, one would expect, as a general rule, that where the same word is used in different provisions of the same Act, the same meaning is intended, unless a contrary intention appears.

The meaning of “agent” in the general law

The word “agent” is a legal term with a well-recognised meaning. As it is put in Bowstead & Reynolds on Agency, 22^nd ed (2021), para 1-001:

“Agency is the fiduciary relationship which exists between two persons, one of whom expressly or impliedly manifests assent that the other should act on his behalf so as to affect his legal relations with third parties, and the other of whom similarly manifests assent so to act or so acts pursuant to the manifestation”.

Professor Francis Reynolds explains:

“An obvious meaning of the word ‘agent’ is ‘someone who acts for or on behalf of another’. … The paradigm reasoning, on which much of the rest of agency law is based, is that one person, usually called the principal, can give authority to, or authorize, another, the agent, to act on his behalf; and that the giving of authority confers on the agent a power to affect the legal position of the person who gave the authority.” (A Burrows (ed), Principles of English Commercial Law, 2015, para 1.01).

As it was put by the High Court of Australia (Dixon, Fullagar and Kitto JJ) in Petersen v Moloney (1951) 84 CLR 91, 94, an agent is “a person who is able, by virtue of the authority conferred upon him, to create or affect legal rights and duties as between another person, who is called his principal, and third parties.”

The extent to which a person is an agent for another depends on the things he is authorised to do for that other. In the Petersen case, for example, an estate agent was authorised to make a contract for the sale of a property by the vendor but was not authorised to receive payment of the price on behalf of the vendor. The estate agent was the vendor’s agent in the legal sense of the word to make the contract on his behalf, so that the vendor was bound by it, but was not his agent to receive the price. The vendor was therefore not bound by the act of estate agent in taking the price money from the purchaser.

This can be contrasted with a case where a person is authorised to receive money on behalf of another and is therefore that other’s agent for the purposes of receipt of the money so that the agent’s receipt is regarded as receipt by the principal. As it is put in the old Latin phrase employed in agency cases, “qui facit per alium facit per se” (he who does something by another does it himself).

So, for example, in Thorne v Heard [1894] 1 Ch 599, CA, the defendants, who were the first mortgagees of property, employed a solicitor (S) to conduct the sale of the property pursuant to their power of sale. S received the sale monies, paid over to the defendants the outstanding sum due under their mortgage and retained the surplus for himself, falsely telling the defendants that he had the authority of the plaintiff, who held a second mortgage, to do this. This misled the defendants into failing to fulfil their duty as first mortgagee to use the surplus to pay the plaintiff, as second mortgagee. S should have used the surplus to pay what was due under the second mortgage, but he concealed from the plaintiff that he had misappropriated the money by continuing to pay interest on the second mortgage as though it still existed. S became bankrupt and the true facts emerged, whereupon the plaintiff sued the defendants for an account of the sale monies and payment of what was due under the second mortgage. The action was held to be statute-barred according to the law of limitation applicable at the time. The defendants had been innocent of any wrongdoing at the relevant time, which was when S received the sale monies on their behalf.

Lindley LJ observed (p 603), “The plaintiff’s right to be paid by the defendants accrued as soon as they received the purchase-money from the purchaser; and the receipt of that money by [S] was clearly a receipt by the defendants, he being their agent, to receive it for them”; but the relevant fraud and concealment occurred later on and the defendants were not implicated in it. Kay LJ and A.L. Smith LJ gave judgments to similar effect. In analysing the position Kay LJ said (p 609), “[S’s] receipt of the money as their agent and by their direction was equivalent to a receipt by the defendants. It absolved the purchaser, and neither against him nor against the persons entitled to the surplus could they be treated as not having received it.” A.L. Smith LJ said (p 613), “It is true that the defendants, fourteen years before action brought, had in their possession [the surplus monies], for the receipt by their agent, [S], was a receipt by them.” The decision of the Court of Appeal was endorsed by the House of Lords, but without such an extensive discussion of this aspect of the case: [1895] AC 495. Lord Herschell LC implicitly accepted the analysis of the Court of Appeal since he observed (p 502) that as soon as S had accounted to the defendants for the money due to them under the first mortgage and explained to them that he held the surplus for others entitled to it, “all agency on their behalf by [S] in relation to that surplus was absolutely at an end.”

It is clear that an agent may be an employee (or servant, in the old terminology) of the principal or an independent contractor engaged by him: Bowstead & Reynolds, op cit, para 1-034. In both Petersen v Moloney and Thorne v Heard, for example, the agents happened to be independent contractors; there are many examples of independent contractors acting as agents. “An employee may or may not be an agent of the employer”; “An agent may or may not be an independent contractor”: G E Dal Pont, Law of Agency 4^th ed, (2020), paras 291, 221; “An agent may be either a servant … or an independent contractor … On the other hand, an employee [ie in the wider sense of someone engaged to do something for the employer] may be a servant or an independent contractor without being in any respect an agent of his employer, having no authority to establish privity of contract between his employer and third persons or to bind him by any other act in law done on his behalf”: J Salmond and P Winfield, Principles of the Law of Contracts (1927), pp 342-343. In both cases, the extent to which they qualify as an agent capable of doing an act which binds their principal (the person who employs them or engages them to carry out some task) in the principal’s relations with a third party depends upon the extent of the authority actually given to them by the principal or (in the case of ostensible authority) which the principal holds them out as having.

The difference between an employee and an independent contractor is, of course, critical when a question of vicarious liability arises. That is, whether one person can be held liable for a wrong committed by another person where the first has committed no wrong at all. But in the law of agency as regards performance of a person’s own duty, the issue is different. It is whether the acts (or omissions), and relevant states of mind of one person in respect of their legal relations with a third party are to qualify in law as being those of another person, on the grounds that the first person is the agent of the other person, who is their principal. They do so where the principal has authorised the agent to act on their behalf in relation to the third party for those purposes.

Wilsons & Clyde Coal Co Ltd v English [1938] AC 57 is an example of this. As the headnote summarises the ruling in the case, “The provision of a safe system of working in a mine is an obligation of the owner, who, if he appoints an agent to perform it, remains vicariously responsible for the agent’s negligence.” The appellant employer argued that its responsibility did not extend beyond appointing competent subordinates to put in place a safe system. Lord Thankerton identified this argument as fallacious (p 65) because the employer “cannot ‘delegate’ his duty in this sense, though he may appoint someone, as his agent in the discharge of the duty, for whom he will remain responsible under the maxim respondeat superior.” Lord Atkin, Lord Wright and Lord Maugham agreed with Lord Thankerton’s speech.

Lord Macmillan’s speech was to the same effect, and Lord Atkin and Lord Maugham agreed with it as well. Lord Macmillan said (p 75) that the relevant obligation was that of the employer mine-owner and observed:

“He cannot divest himself of this duty, though he may – and, if it involves technical management and he is not himself technically qualified, must – perform it through the agency of an employee. It remains the owner’s obligation, and the agent whom the owner appoints to perform it performs it on the owner’s behalf. The owner remains vicariously responsible for the negligence of the person whom he has appointed to perform his obligation for him, and cannot escape liability by merely proving that he has appointed a competent agent. If the owner’s duty has not been performed, no matter how competent the agent selected by the owner to perform it for him, the owner is responsible.”

Lord Wright said (p 80) that “the employer’s obligation … is personal to the employer, and one to be performed by the employer per se or per alios” (by himself or by others, which is to say, by his agents). He explained (pp 83-84):

“There is perhaps a risk of confusion if we speak of the duty as one which can, or cannot, be delegated. The true question is, What is the extent of the duty attaching to the employer? Such a duty is the employer’s personal duty, whether he performs or can perform it himself, or whether he does not perform it or cannot perform it save by servants or agents.”

Lord Maugham set out the same analysis: p 88. The duty to provide a safe system of working and safe machinery is personal to the employer, who “can, and often he must, perform this duty by the employment of an agent who acts on his behalf; but he then remains liable to the employees unless the agent has himself used due care and skill in carrying out the employer’s duty”.

The case law on section 26 of the 1939 Act

In support of its submission based on section 37(1) of the Limitation Act, Primeo relies before the Board, as it did before the Court of Appeal, upon two decisions of the English Court of Appeal concerned with section 26 of the 1939 Act. Both cases were concerned with fraudulent concealment under section 26(b).

The first case, Applegate v Moss [1971] 1 QB 406, concerned a claim against the developer of a housing estate (Mr Moss), brought by persons who had purchased houses to be constructed on the estate from the developer and subsequently discovered that the foundations had not been built in accordance with their contracts with the developer for their construction. Under the relevant contracts, the developer himself promised to construct the houses. It was his obligation to do so, and to do so according to proper standards, albeit it was contemplated that he would engage a builder (Mr Piper) as an independent contractor to carry out the work on his behalf. By the time the defects in the foundations were discovered the ordinary limitation period for bringing a claim had expired. In these circumstances, the purchasers relied on section 26(b) of the 1939 Act and argued that their rights of action had been concealed by the fraud of the developer or his agent, namely the builder whom he had employed as a sub-contractor. The judge found that both the developer and the builder as his agent had concealed the right of action by fraud. That finding was upheld on appeal.

Since the judge’s finding that there had been fraudulent concealment by the developer was upheld, what was said about the position of the builder was not critical to the decision, albeit Edmund Davies LJ gave this as his main reason for dismissing the appeal. He and Lord Denning MR addressed the point. Lord Denning stated (p 413):

“… it is quite plain that Mr Piper, the builder, was the ‘agent’ of Mr Moss. Mr Moss employed Mr Piper to carry out the building work. Mr Piper did the work extremely badly. He was guilty of gross neglect in mixing the concrete. He covered up his disgraceful work. Even if Mr Moss knew nothing about it, nevertheless he must take responsibility for the conduct of Mr Piper.”

Edmund Davies LJ (p 415) recorded the submission of the developer that the word “agent” in section 26(b) had to be given a narrow definition “of somebody acting other than as an independent contractor”, rejected it, and stated (p 415):

“I think the word ‘agent’ as here used embraces an independent contractor. The defendant contracted directly with the purchasers that he would, for the consideration of £1,900, erect ‘a dwelling-house in accordance with the plans and specification hereto annexed.’ In my judgment, it does not lie in his mouth to say that what was done by his builder was not done by his ‘agent’ within the meaning of section 26(b)” (emphasis in original).

The reasoning in Applegate v Moss is based on an entirely conventional approach to the concept of agency in this context. Mr Moss engaged and authorised Mr Piper to act on his behalf to discharge his contractual obligations to the plaintiffs. The decision indicates that the word “agent” in the 1939 Act is not confined to persons who are authorised to make a contract for their principal but extends to persons who are authorised to affect the legal relations of their principal with third parties in other ways, which accords with the conventional legal meaning of the term: see para 250 above. It was the developer who was advancing a submission in favour of a special meaning for the word “agent”, by seeking to exclude all cases of independent contractors. But as explained above, an independent contractor may be an agent. There was no warrant in the language, purpose or context of section 26 of the 1939 Act to justify such an artificial restriction on the meaning of the word “agent”.

This was the point made by Lord Denning and Edmund Davies LJ. Mr Gillis submits that they improperly identified all independent contractors as agents, but they did not. They did not lay down as a proposition that the word “agent” in section 26 covers all independent contractors. Rather, they focused, correctly, on whether in the context of the performance of the developer’s contractual obligations the builder acted as his agent and held that he did. Lord Denning said that Mr Moss employed Mr Piper to carry out the building work, meaning the building work which Mr Moss was contractually obliged to perform. Edmund Davies LJ emphasised that the relevant obligation to be performed was that of Mr Moss, hence his italicisation of the word “he” in the quotation above. Edmund Davies LJ said only that the word “agent” “embraces” an independent contractor, ie is capable of covering an independent contractor in an appropriate case. Since the relevant obligation to build was that of the developer, Mr Moss, he discharged it in the circumstances of the case by engaging someone to perform it on his behalf, as his agent. The reasoning of Lord Denning and Edmund Davies LJ is in line with that in the speeches in the House of Lords in Wilsons & Clyde Coal Co v English, above.

Where a contracting party has a duty to perform, like the developer in Applegate v Moss, and chooses to render performance not by their own person but by engaging another to perform on their behalf, that other person acts as their agent for the purposes of the performance so rendered. Where there is fraudulent concealment by that agent in or about the performance so rendered, then by virtue of section 26(b) that qualifies as the fraudulent concealment of the principal who is the contracting party. This is in line with the position under section 26(a): if a principal authorises a person to act on their behalf and in so acting that person commits a fraud, that fraud is attributable to the principal for the purposes of civil responsibility.

The second case, King v Victor Parsons & Co [1973] 1 WLR 29, concerned similar facts, and was decided by a bench including two of the judges who had sat in Applegate v Moss. The reasoning in that case was not challenged (p 36). As in that case, it was found that there had been fraudulent concealment by the developers themselves. Lord Denning MR adverted to the position of agents (p 34):

“… when I speak of the defendant, I include, of course, his agent: for the statute expressly mentions him. If a defendant employs a contractor to do something for him - and one or other knows that it may well be a wrong or a breach of contract - and keeps quiet about it, then the right of action is concealed by fraud and the defendant cannot avail himself of the statute.”

Megaw LJ and Brabin J applied the reasoning in Applegate v Moss.

There is nothing in these decisions which conflicts with the Board’s interpretation of the word “agent” in section 37(1) of the Limitation Act. They support Primeo’s submissions on this issue.

However, the Court of Appeal expressed its conclusion on this issue in a potentially confusing way. It summarised Primeo’s submission as being that the term “agent” is not limited to those who are given authority to contract on behalf of a principal with third parties but has a more extended meaning and includes independent contractors. As explained above, this submission is correct and is in accordance with the conventional legal meaning of “agent” so long as it is understood that the reference to an independent contractor is to an independent contractor acting as an agent within the scope of the authority conferred by its principal. But in expressing its conclusion the Court of Appeal appeared to treat the concepts of agent and independent contractor as mutually exclusive (rather than as overlapping) and to say that the word “agent” in the statute covered both of them. This gave the impression that the Court of Appeal thought that the two English decisions stand as authority for the proposition that meaning of “agent” in section 26(a) and (b) of the 1939 Act and hence in section 37(1) of the Limitation Act was different from, and wider than, its conventional legal meaning. Mr Gillis criticised their judgment on this basis. The Board is not convinced that this is what the Court of Appeal meant to say, but to the extent they did intend to say this their reasoning cannot be supported.

HSBC’s relevant obligations under the 1996 Custodian Agreement

Under the 1996 Custodian Agreement HSBC agreed to act as custodian of Primeo’s assets held by Primeo in its investment fund. However, until the 2002 Sub-Custody Agreement BLMIS acted as sole custodian without HSBC in fact assuming that role. As the judge found (para 75), the making of the 2002 Sub-Custody Agreement reflected a new tripartite arrangement according to which BLMIS ceased to hold the assets credited to the relevant account as custodian for Primeo and thereafter would hold them as sub-custodian for HSBC. The restructuring of the custodian arrangements was made for various commercial reasons analysed by the judge. From the perspective of Bank Austria and its clients, it was advantageous to confirm that HSBC was performing a real role as custodian since that would remove regulatory doubts about the marketing of the Primeo fund to customers in Austria (para 77). The restructuring meant that the 1996 Custodian Agreement thenceforward applied with full effect according to its terms.

Under clause 9(A) of the Custodian Agreement HSBC was required to hold the securities of Primeo, that is, keep them safe, by physical possession of relevant documents or in book entries and was required to identify securities held by it as being held for Primeo’s account. Under clause 9(C) HSBC was required to keep such books, records and statements as might be necessary to give a complete record of all cash and securities held and transactions carried out by it on behalf of Primeo and permit Primeo to inspect these.

Clause 16(B) provided, in relevant part, as follows:

“In performing its duties hereunder the Custodian [HSBC] may appoint such agents, sub-custodians and delegates as it might think fit to perform in whole or in part any of its duties and discretions (including in such appointment powers of sub-delegation)”.

The clause also imposed other obligations on HSBC, considered in paras 356-371 below, which are not relevant to Primeo’s strict liability claim.

In exercise of its power of delegation under clause 16(B), HSBC entrusted the performance of its duties as custodian to BLMIS. This in no way relieved HSBC from its obligations under the 1996 Custody Agreement, which is why it is accepted that HSBC is strictly liable under that agreement for breach of its safekeeping obligation by reason of BLMIS’s theft of Primeo’s assets purportedly held by BLMIS as sub-custodian for HSBC.

BLMIS’s relevant obligations under the 2002 Sub-Custody Agreement

The authority which HSBC gave to BLMIS to act on its behalf in relation to the safekeeping of the assets allocated to Primeo is primarily set out in the 2002 Sub-Custody Agreement between HSBC and BLMIS. Under that agreement, HSBC appointed BLMIS to hold the property of which HSBC was custodian on behalf of specified customers, including Primeo, in safe custody, and/or to administer it on the terms and conditions set out in the agreement. In particular, clause 3.4 required BLMIS to carry out HSBC’s instructions. Clause 4 directed BLMIS to open and maintain an account for the exclusive benefit of each of the specified customers, in the name of “HSBC ‘Special Custody Account for [the name of the specified customer]’”. Clause 5.1 required BLMIS to hold in the account all property acquired by it in pursuance of the agreement. Clause 5.2 provided:

“All Property required hereunder to be held in the Account shall be physically segregated from the general assets of the Sub-Custodian … and the Sub-Custodian shall mark its records so as to identify Property as Property held in a fiduciary capacity to the order of the Bank.”

Clause 8 required BLMIS to keep a complete record of all property held and/or administered by it and “transactions carried out by it on behalf of the customers of the Bank”. Clause 12.1(a) required BLMIS to “ensure there is legal separation of non-cash assets held under custody and that such assets are held on a fiduciary basis”. Clause 12.1(d) required BLMIS to “furnish to the Bank on an annual basis or as otherwise requested by the Bank confirmation from the Sub-Custodian that non-cash assets are held by them on a fiduciary basis”. Clause 12.2 required BLMIS to use its best efforts to ensure that the Bank was in compliance with all regulatory legislation to which it was subject as custodian for its customers. Other provisions imposed obligations on BLMIS in relation to reporting and confidentiality, among other matters. Clause 16.3 provided that the agreement was not assignable by either party without the written consent of the other.

The question which arises is, in what capacity did BLMIS act when purporting to discharge its obligations under the 2002 Sub-Custody Agreement? When acting as sub-custodian of Primeo’s assets, maintaining records of the assets held, receiving payments from Primeo by way of further investments and giving receipts therefor, so as (apparently) to satisfy the obligations of HSBC to do these things under the 1996 Custodian Agreement, did BLMIS act as HSBC’s agent within the meaning of section 37(1) of the Limitation Act?

Analysis

Mr Smith submits that when BLMIS purported to comply with the 2002 Sub-Custody Agreement it acted as HSBC’s agent to perform its obligations owed to Primeo under the 1996 Custodian Agreement. In the Board’s view, this is clearly correct. Those obligations were personal to HSBC and required HSBC (among other things) to safeguard Primeo’s assets. That is an obligation which required performance by HSBC. HSBC had not undertaken merely to stand as a surety for the due performance of duties owed to Primeo by BLMIS. Under the restructured tripartite custody arrangement BLMIS owed no such duties. In order to perform its obligation, HSBC had to act by its agents, either in the form of its own employees or in the form of independent contractors appointed to act as its agents for this purpose. BLMIS was HSBC’s agent in carrying out the safeguarding obligation of HSBC.

This analysis is confirmed by the terms of clause 16(B) of the 1996 Custodian Agreement, set out at para 272 above. That states that “[i]n performing its duties” under the agreement HSBC may appoint such agents, sub-custodians and delegates “as it might think fit to perform in whole or in part any of its duties …” (emphasis added). If such persons were appointed to perform HSBC’s duties, they did so as its agents. The use of the other terms alongside “agent” does not indicate otherwise, since the parties to an agreement cannot by the use of labels convert what is in law an agency arrangement into something which is not. In any event, each of the terms used is consistent with the person appointed acting as HSBC’s agent in relation to some aspect or other of HSBC’s own duties.

The actions and omissions of BLMIS pursuant to the 2002 Sub-Custody Agreement as regards performance of HSBC’s obligations under the 1996 Custody Agreement plainly affected HSBC’s legal relations with a third party, namely Primeo. Therefore BLMIS falls within the concept of an agent under the general law: para 250 above.

If HSBC took no steps itself to safeguard the assets for which it was designated custodian and appointed no one to take such steps on its behalf, it would have been in breach of the 1996 Custodian Agreement on grounds of non-performance. Its appointment of BLMIS to undertake the safeguarding on its behalf was intended to achieve the due discharge of HSBC’s obligations vis-à-vis Primeo through performance by BLMIS on its behalf, and to the extent that BLMIS did what it was supposed to do its actions would have achieved that result. Conversely, BLMIS’s failure to ensure that the assets were safeguarded had the effect that HSBC was placed in breach of its safeguarding obligation owed to Primeo. That is the reason why the strict liability claim is good in law. As Mr Smith submitted, BLMIS was authorised by HSBC to act on its behalf to perform its duties under the 1996 Custodian Agreement with Primeo, and BLMIS’s acts in that regard would affect HSBC’s legal relations with Primeo either by discharging its contractual duties to Primeo or, if such duties were not discharged as required by the terms of the contract, by exposing it to liability to Primeo for breach of duty.

It may be added that under the tripartite arrangement BLMIS acted as HSBC’s agent by affecting HSBC’s legal relations with Primeo in other respects as well. When Primeo paid money to HSBC to be invested in assets to be held in safekeeping for its account, BLMIS received those monies as agent for HSBC for the purposes of creating rights and obligations as between Primeo and HSBC under the 1996 Custodian Agreement. If any dispute arose, HSBC could not have denied that the monies were covered by its obligations under that agreement, because BLMIS received them as its agent: see the judgments in the Court of Appeal in Thorne v Heard (above). Similarly, if and when BLMIS provided statements to Primeo of the assets which were held to Primeo’s account under the tripartite arrangement, HSBC was bound by those statements because they were provided by BLMIS as its agent to fulfil HSBC’s obligations under the 1996 Custodian Agreement, including for example under clause 9(C).

Conclusion

Applying section 37(1) in accordance with its clear and ordinary meaning, it is plain that BLMIS deliberately concealed relevant facts when acting as HSBC’s agent in rendering performance to satisfy HSBC’s obligations under the 1996 Custodian Agreement. The running of time was therefore postponed pursuant to section 37(1) of the Limitation Act. HSBC’s appeal on this point should therefore be dismissed.

Contributory Negligence

In these proceedings the respondents have maintained that any award of damages to Primeo falls to be reduced by reason of its contributory negligence.

At first instance the judge held that the defence of contributory negligence applies to Primeo’s claim against Bank of Bermuda in respect of the NAV calculations because the claim is based on an implied term in the Administration Agreement to exercise reasonable care and skill which is co-extensive with a tortious duty of care. However, the judge held that the defence does not apply to Primeo’s claims against HSBC for breach of clause 16(B) of the Custodian Agreement because this did not give rise to an implied term which would be co-extensive with a tortious duty to take care. With regard to the claim against Bank of Bermuda, the judge concluded that Primeo was, to a very substantial degree, the author of its own misfortune and he would, accordingly, have reduced any award of damages against Bank of Bermuda by 75% on grounds of contributory negligence.

On appeal, the Court of Appeal considered that both respondents, in respect of their fault-based liability, could rely upon Primeo’s contributory negligence. In the case of HSBC it came to this conclusion on the basis that the indemnity provisions of clause 16 of the Custodian Agreement meant that unless HSBC or its agent was negligent or in wilful breach of contract HSBC was indemnified from all liabilities, and accordingly the duty was in reality a duty to exercise reasonable skill and care. On this basis there existed concurrent and co-extensive duties in contract and in the tort of negligence. The Court of Appeal further considered that the appropriate reduction for contributory negligence in the case of the claims against both respondents was 50%.

On this appeal Primeo submits that the Court of Appeal was wrong to hold that the relevant claims of Primeo against the respondents should be reduced for contributory negligence. In particular, Primeo submits that:

Primeo no longer pursues its case that it was wrong for the courts below to conclude that Primeo could be guilty of contributory negligence in circumstances where Primeo had instructed the respondents, reputable, experienced and specialist service-providers, to look after matters of administration and custody, and had trusted them to carry out their contractual duties. However, Primeo says that this aspect of the case is relevant to the question of the extent of its responsibility for the loss it suffered and maintains that the judge failed to take it properly into account when he determined that the damages claimed against Bank of Bermuda should have been reduced by 75%.

The respondents take issue on these points and cross appeal against the decision of the Court of Appeal to reduce the appropriate reduction for contributory negligence in relation to Bank of Bermuda from 75% to 50%. In addition, the respondents argued that the Court of Appeal did not specifically address the level of reduction as against HSBC, but that the judge’s findings of Primeo’s degree of contributory negligence should be given effect by ordering the same reduction of 75% as against both respondents.

Contributory negligence as a defence to contract

Statutory provisions

The relevant statutory provisions were first enacted in the Cayman Islands by the Law Reform (Contributory Negligence) Law 1964 (Law 9 of 1965) (“the 1964 Law”). In 1977 the 1964 Law was revised pursuant to the Law Revision Act 1975 (Law 19 of 1975) (“the Law Revision Law”). As part of this exercise, it was consolidated with the Fatal Accidents Law (Cap 54) and the Law Reform (Tortfeasors) Law 1964 (Law 8 of 1965) and reissued under the title of the Law of Torts Reform Law (Revised) (“the 1977 Revision”). The current version, applicable in this case, is the Torts (Reform) Law (“the 1996 Revision”).

Section 8(1) of the 1996 Revision provides:

“Where any person suffers damage as the result partly of his own fault and partly of the fault of any other person or persons, a claim in respect of that damage shall not be defeated by reason of the fault of the person suffering the damage, but the damages recoverable in respect thereof shall be reduced to such extent as the court thinks just and equitable having regard to the claimant’s share in the responsibility for the damage:

Provided that –

(a)

this subsection shall not operate to defeat any defence arising under a contract; and

(b)

where any contract or enactment providing for the limitation of liability is applicable to the claim, the amount of damages recoverable by the claimant by virtue of this subsection shall not exceed the maximum limit so applicable.”

Section 8(1) of the 1996 Revision is materially identical to section 3(1) of the 1964 Law and section 8(1) of the 1977 Revision.

“Fault” was defined as follows in the 1964 Law:

“‘fault’ means negligence, breach of statutory duty or other act or omission which gives rise to a liability in tort or would, apart from this Law, give rise to the defence of contributory negligence.”

Section 2 of the 1977 Revision and the 1996 Revision, however, defined “fault” as follows:

“‘fault’ means an act creating a liability in tort [or] which, prior to the operation of this Law, would have given rise to the defence of contributory negligence;”

It was common ground before us that the word “or” had been inadvertently omitted from the definition in the statute during the revision process in 1977, as was explained by Jones J at para 314 of his judgment. Under section 3 of the Law Revision Law the process of revision does not permit any substantive amendment of the law. As a result, the revised version must be interpreted as if the accidental omission had not occurred.

Section 8(1) of the 1996 Revision is materially identical to section 1(1) of the 1945 Act. However, “fault” is defined in section 4 of the 1945 Act as follows:

“‘fault’ means negligence, breach of statutory duty or other act or omission which gives rise to a liability in tort or would, apart from this Act, give rise to the defence of contributory negligence.”

This definition, as the Board has shown, was adopted by the 1964 Law.

Authorities

In Sayers v Harlow Urban District Council [1958] 1 WLR 623 the plaintiff suffered a personal injury when attempting to escape from a public lavatory operated by the defendants in which she had been trapped. She had paid to use the lavatory and she claimed that “the damage she suffered was due to the fault of the defendants, that fault being in the form of breach of the duty of care owed to her, whether or not arising under the implied contract when she made use of the lavatory”. The Court of Appeal held that she was entitled to recover but reduced her damages by 25% for contributory negligence. The judgments do not include any consideration of the applicability of the 1945 Act. Lord Evershed MR observed (at p 625) that nothing turned upon the foundation of liability.

In Basildon District Council v J E Lesser (Properties) Ltd [1985] QB 839 Judge John Newey QC held that contributory negligence was not available as a defence to a claim in contract where the third defendant had promised under the terms of a deed to indemnify the plaintiffs for loss or damage occasioned by the second defendant’s performance of a building contract. The judge considered that contributory negligence had developed at common law as a defence to an action founded on fault. It later came to be applied to other, older, stricter forms of liability in tort, such as those based on status. The 1945 Act had retained the concept of blameworthiness but in contract blameworthiness was irrelevant. In the claim against the third defendant there was no room for contributory negligence, although in the assessment of damages causation and the plaintiff’s duty to mitigate his loss were very relevant. Whether the 1945 Act changed the common law so as to make contributory negligence a ground for reducing damages in contract must depend upon its wording but it would be surprising if Parliament when limiting the effect of contributory negligence in tort introduced it into contract. In considering the judge’s approach, it should be noted, however, that, employing the classification later developed by Hobhouse J in Vesta v Butcher which is discussed below, this was a category 1 case and that with regard to category 3 cases Judge Newey (at pp 847-849) indicated his approval of the approach of Pritchard J in the New Zealand case of Rowe v Turner Hopkins, which is considered below. (See also Vesta v Butcher per O’Connor LJ at p 865 D-E.)

AB Marintrans v Comet Shipping Co Ltd [1985] 1 WLR 1270 concerned a charterparty dispute over responsibility for the consequences of the shifting of a cargo which had been stowed in a manner which rendered the vessel unseaworthy. The owners claimed the balance of the charterparty hire. The charterers had made deductions for a period off-hire which the shipowners maintained had been caused by the fault of stevedores who, under the terms of the charterparty, were persons for whom the charterers were responsible. The arbitrators found that the relevant delay was predominantly the fault of the master, the owners’ servant, but had been contributed to by the fault of the stevedores, and they apportioned blame under the 1945 Act accordingly. Neill LJ, sitting as a judge of the Commercial Court, held that section 1 of the 1945 Act did not apply so as to permit the damages recoverable by a plaintiff in respect of a contractual or other non-tortious claim to be reduced by reason of the plaintiff’s contributory negligence, even though the breach of contract relied upon was in the nature of a breach of a contractual duty of care. On behalf of the owners, it was submitted that while the 1945 Act did not apply in a situation where the claim was in respect of some strict or absolute obligation in contract, it did apply and an apportionment could therefore be made where the parties owed each other a duty of care in contract or could on the facts have made a claim in tort for breach of a duty of care. Neill LJ considered, however, that the 1945 Act was applicable in cases where, “if the Act had not been passed, a claim of the nature specified would have been defeated or would be likely to have been defeated by reason of the fault of the person suffering the damage.” Moreover, it was common ground that this “fault” would have been “an act or omission which … would, apart from this Act, give rise to the defence of contributory negligence.” Once it was conceded that there was no defence of contributory negligence before the 1945 Act where the claim against the defendant lay in contract, it followed that if the claim was in contract there could be no relevant “fault of the person suffering the damage” within the meaning of that Act (pp 1287H–1288A). The judge added (p 1288G-H):

“I appreciate that the construction which I have adopted may well lead in some cases to unsatisfactory results. Thus it may be that a plaintiff will be able to avoid the apportionment provisions by suing in contract when a claim in tort would be as or more appropriate. But this is not a problem for a judge at first instance to attempt to solve by placing a strained construction on a statute. The matter may, however, be a suitable topic for study by those with responsibilities for law reform. Indeed, I see great force in the contention that the same rule should apply to claims whether they are based in contract or tort where the act complained of involves the breach of a duty of care.”

In Vesta v Butcher the plaintiffs, a Norwegian insurance company, had insured the owners of a fish farm against loss of fish from any cause whatsoever. The plaintiffs reinsured with London underwriters through brokers for 90% of the risk. It was a condition of the insurance and reinsurance policies that a 24-hour watch be kept on the farm. The owners notified the plaintiffs that they could not comply with the 24-hour watch condition. The plaintiffs told the brokers that there would be no 24-hour watch and that the clause should be removed. The brokers did not implement the request and the plaintiffs did not follow it up. A storm damaged the farm causing a loss of fish at a time when the farm was not being watched. The plaintiffs settled the owner’s insurance claim and sought indemnity under their reinsurance policy. The reinsurers repudiated liability on the ground, inter alia, that there had been a breach of the 24-hour watch condition. The plaintiffs sued the reinsurers to recover the indemnity and, in the alternative, sued the brokers for damages for breach of duty. The brokers contended that even if they had been negligent, the plaintiffs’ failure to follow up the matter amounted to contributory negligence. The plaintiffs contended in reply that the brokers could not rely on a defence of contributory negligence because the 1945 Act did not apply to a claim for breach of contract.

At first instance ([1986] 2 All ER 488) Hobhouse J held that the reinsurers were not entitled to repudiate the reinsurance policy. Nevertheless, he set out (at pp 507ff) his obiter views on the availability of a defence of contributory negligence to the claim against the brokers. The plaintiffs’ pleaded case was that there were concurrent contractual and tortious duties, breach of which could be put either as a claim in contract for breach of an implied term to exercise reasonable skill and care or as a claim in the tort of negligence. The plaintiffs argued that although the content of the obligation and breach was the same in either case, the existence of a right to formulate the claim in contract took the claim out of the ambit of the 1945 Act. In order to examine this submission, the judge identified three categories of case:

This classification has been employed in many of the subsequent cases. According to the judge’s analysis, a defence of contributory negligence under the 1945 Act leading to a reduction in the damages payable is available in category 3 cases but not in cases within category 1 or category 2.

Hobhouse J observed that the category 3 question had arisen in very many different types of case and that the answer was treated as so obvious as to pass without comment:

“It is commonplace that actions are brought by persons who have suffered personal injuries as a result of the negligence of the person sued and that there is a contractual as well as tortious relationship. In such cases apportionment of blame is invariably adopted by the court notwithstanding that the plaintiff could sue in contract as well as in tort.”

Sayers v Harlow UDC was such a case. He considered that the power to make an apportionment was part of the ratio decidendi of the Court of Appeal’s decision and was binding on him. In his view there were innumerable similar decisions to the same effect, very many by appellate courts, and the “plaintiffs have not had the temerity to suggest that they are all wrong”. However, he accepted that Judge Newey in Basildon DC and Neill LJ in Marintrans had implicitly said that they were wrongly decided. Hobhouse J considered that Marintrans was probably a category 1 case but, failing that, not more than a category 2 case. Nevertheless, he accepted (at p 509J) that the reasoning of Neill LJ in Marintrans expressly covered any case where a contractual cause of action could be formulated even though there was a fully concurrent tortious cause of action. Hobhouse J considered (at p 510 E-F) that the 1945 Act was not concerned with displacing contractual rights, as was demonstrated by the provisos in section 1(1), but those provisos contemplated that the Act was capable of application where there was a contractual relationship between the parties and thus cast doubt on the approach adopted by Neill LJ.

Hobhouse J concluded that where, as in that case and in any similar category 3 case, there was no express contractual provision which defined the parties’ rights and liabilities in a different way, apportionment of blame and liability was open to the tribunal:

“… the correct analysis is that where there is independently of contract a status or common law relationship which exists between the parties and which can then give rise to tortious liabilities which fall to be adjusted in accordance with the 1945 Act, the relevant question in any given case is whether the parties have by their contract varied that position. Here they patently have not.” (at p 510G)

“If the contract does not on its true construction disclose an intention to redefine or vary in any of these ways the legal incidents of the common law relationship that exists, those incidents remain. Apportionment of blame, and therefore of liability, has since 1945 been one of these incidents.” (at p 510J)

On appeal, the Court of Appeal ([1989] 1 AC 852) agreed with Hobhouse J that the reinsurers were liable under the contract of reinsurance and so the brokers’ negligence had caused no loss. As a result, once again, the observations of the Court of Appeal on the availability of contributory negligence were obiter. However, all three members of the Court of Appeal agreed with Hobhouse J that a defence of contributory negligence was in principle available.

O’Connor LJ held that the 1945 Act applied in that case for essentially the same reasons as Hobhouse J had given at first instance. O’Connor LJ noted that the claim against the brokers was pleaded in contract and tort. This he saw as but a recognition of a clearly established principle that where under the general law a person owes a duty to another to exercise reasonable care and skill in some activity, a breach of that duty gives rise to a claim in tort notwithstanding the fact that the activity is the subject matter of a contract between them. In such a case the breach of duty would also be a breach of contract. Since the decision of the House of Lords in Hedley Byrne & Co Ltd v Heller & Partners Ltd [1964] AC 465 the relationship between the brokers and the plaintiffs had been an example of such a case. He then referred to the submission on behalf of the plaintiffs that, nevertheless, if a plaintiff makes his claim in contract contributory negligence cannot be relied on by the defendant, whereas it is available if the claim is made in tort, and observed:

“If this contention is sound then the law has been sadly adrift for a very long time for it would mean that in employers’ liability cases an injured employee could debar the employer from relying on any contributory negligence by framing his action in contract.” (at p 860D)

O’Connor LJ rejected the reasoning of Neill LJ in Marintrans and his conclusion in that case that where a claim was made in contract there was no power under the 1945 Act to apportion even in cases where the breach complained of was co-extensive with a breach of duty giving rise to a liability in tort. In concluding that there was a power to apportion in a category 3 case, O’Connor LJ adopted the reasoning of Prichard J in relation to the equivalent New Zealand provision in Rowe v Turner Hopkins, which is considered in detail below.

Neill LJ agreed, although with some hesitation. Referring to his judgment in Marintrans, he found it difficult to assent to the proposition that on its proper construction the 1945 Act could be applied so as to reduce a plaintiff’s claim where the action was brought for damages for breach of contract. However, having read the draft judgment of O’Connor LJ and having considered the cases there referred to, in particular the judgment of Prichard J in Rowe v Turner Hopkins, he felt bound to concur. Employing “the valuable classification” of Hobhouse J, Vesta v Butcher was clearly a category 3 case where the brokers’ liability in contract was the same as their liability would have been in tort. Accordingly, he agreed that as the claim against the brokers could have been framed as a breach of a duty of care in tort, any damages awarded to the plaintiffs could properly be reduced and apportioned in accordance with the 1945 Act.

Sir Roger Ormrod agreed that the defence of contributory negligence was available but he seems to have regarded the plaintiffs’ claim as a claim in tort rather than in contract. He stated that he remained unconvinced that contributory negligence at common law had any relevance in a claim in contract and observed:

“Had contributory negligence been a defence at common law to a claim for damages for breach of contract the reports and the textbooks prior to 1945 would have been full of references to it.” (at p 879 B-C)

Citing Prichard J in Rowe v Turner Hopkins, he considered that the context of the 1945 Act and the language of section 1 made clear that the Act is concerned only with tortious liability and that the power to apportion only arises where the defendant is liable in tort, and that concurrent liability in contract, if any, is immaterial. The 1945 Act only applies in a case in which, but for the Act, the claim for damages would have been “defeated by reason of the fault of the person suffering the damage”. While the brokers’ liability could be pleaded as a breach of an implied term, it might be more accurate to say that the existence of the contract created a degree of proximity between the plaintiff and the brokers sufficient to give rise to a duty of care and, therefore, to a claim in negligence (at p 779E-F).

In Barclays Bank PLC v Fairclough Building Ltd [1995] QB 214 the plaintiff sued for damages for breach of two obligations under a standard term building contract, both of which required strict performance. The judge held the defendant liable but also held that the plaintiff had failed to supervise the work so as to prevent the defendant’s breaches, with the result that the award of damages should be reduced by 40% because of the plaintiff’s contributory negligence.

On appeal, Beldam LJ, adopting the Hobhouse J classification and observing that the case before him was not a category 3 case, said with regard to category 3 cases:

“To regard the definition of fault in section 4 as extending to cases such as employer’s liability places no great strain on the construction of the words used. In 1945 actions brought by an employee whether framed in contract or tort were usually regarded as actions in negligence and the defence of contributory negligence was by no means uncommon.” (at p 229B)

(See, also, O’Connor LJ in Vesta v Butcher at p 860D cited at para 300 above.)

Beldam LJ considered, however, that so far as category 1 cases were concerned, such as that before the court, there was no decision in which contributory negligence had been held to be a defence and there were powerful dicta to the contrary effect (at p 229C-D). He concluded (at p 230F) that contributory negligence was not a defence to a claim for damages founded on breach of a strict contractual obligation.

Simon Brown LJ accepted the view of Hobhouse J in Vesta v Butcher that apportionment of blame and liability was open to the court in any ordinary category 3 case unless the parties by their contract had varied that position, because the contract in such cases added nothing to the common law position. However, when, as in a category 1 case, the contractual liability was by no means immaterial, when rather it was a strict liability arising independently of any negligence on the defendant’s part, there were compelling reasons why the contract, even assuming it to be silent as to apportionment, should be construed as excluding the operation of the 1945 Act. The very imposition of a strict liability on the defendant was inconsistent with an apportionment of the loss (at p 233B-F).

Nourse LJ, agreeing with Beldam LJ, observed (at p 234C) that it ought to have been perfectly obvious that the 1945 Act was never intended to obtrude the defence of contributory negligence into an area of the law where it had no business to be.

In Standard Chartered Bank v Pakistan National Shipping Corpn (Nos 2 and 4) [2003] 1 AC 959 (“Standard Chartered Bank”) the House of Lords held unanimously that a defence of contributory negligence under the 1945 Act could not apply to a claim in deceit. Similarly, in Pritchard v Co-operative Group Ltd [2011] EWCA Civ 329; [2012] QB 320 (“Co-operative Group v Pritchard”) the Court of Appeal held that a defence of contributory negligence under the 1945 Act could not apply to a claim in assault and battery. These decisions are considered in detail at paras 325-329 below.

Commonwealth decisions

New Zealand

Rowe v Turner Hopkins & Partners [1980] 2 NZLR 550 concerned a claim against solicitors for professional negligence. In a passage which has proved highly influential Prichard J observed with regard to the corresponding provisions in the New Zealand Contributory Negligence Act 1947 (sections 2 and 3(1)) (at pp 555-556):

“To my mind, the Act provides its own interpretation if it is acceptable to regard the definition of ‘fault’ in s 2 as comprising two limbs – the first referable to the defendant’s conduct, the second to the plaintiff’s conduct. Section 2 defines ‘fault’ as meaning ‘negligence, breach of statutory duty, or other act or omission which gives rise to a liability in tort’ (the first limb). It then goes on to include any act or omission which ‘would, apart from this Act, give rise to the defence of contributory negligence’ (the second limb). In my view, the first limb of the definition is plainly directed to defining ‘fault’ as it relates to the conduct of the defendant – in other words, as it relates to the plaintiff’s cause of action. This phrase is qualified by the expression ‘which gives rise to a liability in tort’. It follows that no negligence, breach of statutory duty and no other act or omission of the defendant will bring s 3(1) into play unless it is one which gives rise to liability in tort. In other words, the Act applies only when the plaintiff’s cause of action is in respect of some act or omission for which the defendant is liable in tort. Conceivably, the defendant may be concurrently liable in contract – but that is immaterial – the sine qua non is conduct creating liability in tort.

The second limb of the definition is concerned with and is referable only to the conduct of the plaintiff. It relates not to any cause of action but to conduct which, prior to the Act, would give rise to the defence of contributory negligence and which is now to be regarded as that conduct on the part of a plaintiff which will lead not to a complete defence but to a reduction in damages. Before the enactment of the Contributory Negligence Act, the defence of contributory negligence was a complete defence in tort: it was not a defence in contract – where the issue was more likely to be simply causation.

I therefore conclude, in the absence of any clear authority to the contrary, that the first limb of the definition of s 2 determines the meaning of the word ‘fault’ as it relates to the plaintiff’s cause of action: that accordingly, the Contributory Negligence Act cannot apply unless the cause of action is founded on some act or omission on the part of the defendant which gives rise to liability in tort: that if the defendant’s conduct meets that criterion, the Act can apply – whether or not the same conduct is also actionable in contract. By the same token – the second limb of the definition means simply and logically that no act or omission of the plaintiff will entitle the defendant to a reduction of damages unless it amounts to the sort of conduct which, prior to the enactment of the Contributory Negligence Act, would have afforded a defence of contributory negligence.

I turn then to the second question – whether the plaintiff’s cause of action is founded solely in contract or whether it is, concurrently, a claim in tort.”

This passage was cited with approval by Judge Newey QC in Basildon DC v JE Lesser Ltd (at pp 847H-848F, 849D) and by O’Connor LJ and Sir Roger Ormrod in Vesta v Butcher (at pp 865E-866E, 879D). (See also Barclays Bank PLC v Fairclough Building Ltd per Simon Brown LJ at p 233E-F.)

In this way, Prichard J concluded in Rowe v Turner Hopkins that the defence of contributory negligence does apply in category 3 cases. He was, however, bound by authority (McLaren Maycroft & Co v Fletcher Development Co Ltd [1973] 2 NZLR 100 (CA)) to hold that in an action founded on an allegation of professional negligence the duty of care was purely contractual. In these circumstances the New Zealand Contributory Negligence Act 1947 could not apply because the cause of action was not founded on some act or omission on the part of the defendant which gave rise to liability in tort. When the matter went on appeal, the New Zealand Court of Appeal (Cooke, McMullin and Roper JJ, [1982] 1 NZLR 178) held that the judge’s finding of negligence on the part of the defendants was not justified on the facts. As a result, it was not necessary to decide the issue of the scope of the defence of contributory negligence. Cooke and Roper JJ therefore merely referred (at p 181) to the view that the defence “can apply wherever negligence is an essential ingredient of the plaintiff’s cause of action, whatever the source of the duty”. The Board considers this to be a helpful formulation of the principle underlying the requirement of a defendant’s fault for this purpose.

In Mouat v Clark Boyce [1992] 2 NZLR 559, another case of professional negligence by a solicitor, the New Zealand Court of Appeal (Cooke P, Richardson and Gault JJ) held that from the outset of interviewing the plaintiff, the defendant solicitor had owed her a duty of care in tort. There had been a breach of that duty and apportionment was available under the Contributory Negligence Act 1947. It was therefore unnecessary to consider the consequences of a breach of a later contract of retainer or questions of concurrent liability in tort and contract, or of the application of the Act to negligence solely in breach of contract. However, Cooke P observed (at p 564) that when the contract of retainer was made it did not convert the duty into one arising only in contract. At all times it was a duty to take reasonable care of the plaintiff’s interests. The duty stemmed from the relationship, it being immaterial to the extent of the duty that at some moment in the course of the relationship a contract was made, unless the contract limited the duty. He continued:

“Just as, once a retainer has been accepted, the content of the duty of care in such a case is identical whether derived from theoretical sources of tort, contract or equity, or as I think from all of them in a situation where they overlap, so in my opinion its source or sources do not affect the power to apportion.” (at p 564)

Cooke P expressly agreed with the decision in Vesta v Butcher “that the Act does apply in cases of concurrent sources of duty” adding that the English Court of Appeal had apparently been unaware that in Rowe v Turner Hopkins on appeal the New Zealand Court of Appeal, without deciding the point, had drawn attention to the view (different from that expressed by Prichard J at first instance) that the Act can apply “wherever negligence is an essential ingredient of the plaintiff’s cause of action, whatever the source of the duty” (at p 564). While he favoured a wider view that the Act applied both to cases of a purely contractual duty of care and to cases of a duty of care arising from contract and tort concurrently, he also supported the narrower view that “the Act applies wherever there is a duty of care, one source of which is the law of tort, no matter whether or not there is another source of contract or otherwise” (at p 565).

Australia

In Rowe v Turner Hopkins (paras 309–310 above) Prichard J (at p 554) drew particular attention to two Australian authorities in which the view was taken that the local equivalent legislation did not apply to claims in contract. In Belous v Willetts [1970] VR 45, a professional negligence claim against solicitors, Gillard J in the Supreme Court of Victoria held that section 26 of the Wrongs Act 1958 (Victoria), which was in similar terms to the other statutory provisions under consideration here, did not apply in a claim which was held to be in contract only. Similarly, in ASJames Pty Ltd v Duncan, [1970] VR 705 McInerny J in the Supreme Court of Victoria undertook a detailed examination of the issue and held, obiter, that the Wrongs Act did not apply in contract. McInerny J observed, at p 726, “I think it altogether unlikely that the Act was ever intended to apply to any other actions than those founded in tort.”

The issue came before the High Court of Australia in Astley v Austrust Ltd [1999] Lloyd’s Rep PN 758; (2000) 197 CLR 1 (“Astley”), a decision on which Primeo places particular reliance. The plaintiff, a trustee company, sued its solicitors in contract and tort claiming that they had negligently failed to advise it that it should not accept the office of trustee of an existing trading trust without excluding its personal liability for losses. The defendant solicitors denied liability and, in the alternative, alleged that the plaintiff had been contributorily negligent by failing to make its own assessment of the risks posed. In the Supreme Court of South Australia, the trial judge held the defendant liable but decided that the damages awarded should be reduced by 50% for the plaintiff’s negligence pursuant to section 27A of the Wrongs Act 1936 (South Australia), which was in similar terms to the statutory provisions under consideration here. On appeal, the Supreme Court of South Australia allowed the plaintiff’s appeal against the finding of contributory negligence and dismissed the defendants’ appeal against the finding that they had been negligent. On a further appeal, the High Court of Australia held that the plaintiff had been negligent but (Callinan J dissenting) that contributory negligence was not a defence to an action in contract at common law and did not give rise to apportionment of liability under the legislation, notwithstanding that the contractual duty was concurrent with a duty in the tort of negligence.

In a joint judgment the majority (Gleeson CJ, McHugh, Gummow and Hayne JJ) considered that the law had evolved to the conclusion that concurrent liabilities in both contract and tort may arise in cases of professional negligence. Persons who give consideration for the provision of services expect that those services will be provided with due care and skill, and reliance on an implied term giving effect to that expectation should not be defeated by the recognition of a parallel and concurrent obligation under the law of negligence. Given the differing requirements and advantages of each cause of action, there was no justification in recognising the tortious duty to the exclusion of the contractual duty.

Rejecting the reasoning of Prichard J in Rowe v Turner Hopkins, the majority considered that since, as Prichard J accepted, the fault of the plaintiff was never a defence to an action for breach of contract, the direction to apportion made by the legislation was inoperative in cases of contract. In their view, the case law in this area was unsatisfactory, displaying substantial flaws of reasoning and was overall in a state of confusion. Furthermore, the tripartite division adopted by the United Kingdom cases was unacceptable. Those decisions which had applied the 1945 Act to breaches of contract were wrong and should not be followed.

So far as statutory construction was concerned, the majority considered that “on any fair reading of the apportionment legislation against the background of the mischief it was intended to remedy, it is clear to the point of near certainty that the legislation does not and was never intended to apply to contractual claims” (at p 774, col 2). A series of interpretative considerations pointed to the legislation having no purpose of affecting the damages awarded for breach of contract. But if there were any doubt about the matter, the state of the pre-existing law and the purpose of the legislation made it clear that the legislation did not affect actions for breach of contract. At common law, contributory negligence was a complete defence to an action in tort for negligence. “No case can be found in the books where contributory negligence, as such was ever held to be a defence to an action for breach of contract” (at p 775, col 1). In an extended passage the majority addressed the pleading of the defence.

“If contributory negligence was a defence to a breach of contract, it would have had to have been specially pleaded. Yet neither the case law nor the practice works contain any reference to such a plea. It is impossible to accept that contributory negligence, as such, was ever a defence to a claim in contract when neither the cases nor the books of pleading and practice make any reference to it.” (at p 775, col 2)

The fact that contributory negligence was not a defence to an action in contract pointed irresistibly to the conclusion that the apportionment legislation was concerned only with actions in tort and did not affect awards of damages based on breach of contract.

In their judgment the majority observed that the purpose of the enactment of the apportionment legislation had been to allow recovery of damages in cases where contributory negligence would defeat an action in tort. It would be strange if a rule introduced to do away with an absolute defence to a claim in negligence, diminished the rights of a plaintiff who sued in contract.

Turning to policy considerations the majority observed:

“But when the nature of an action for breach of a contractual term to take reasonable care and the nature of an action in tort for breach of a general law duty of care are examined, it is by no means evident that there is anything anomalous or unfair in a plaintiff who sues in contract being outside the scope of the apportionment legislation. Tort obligations are imposed on the parties; contractual obligations are voluntarily assumed.” (p 777, col 1)

They explained that in contract the plaintiff gives consideration which is often very substantial and that the terms of the contract allocate responsibility for the risks of the enterprise including the risk that the damage suffered by one party may arise partly from the failure of that party to take reasonable care for the safety of that person’s property or person:

“Absent some contractual stipulation to the contrary, there is no reason of justice or sound legal policy which should prevent the plaintiff in a case such as the present recovering for all the damage that is causally connected to the defendant’s breach even if the plaintiff’s conduct has contributed to the damage which he or she has suffered. By its own voluntary act, the defendant has accepted an obligation to take reasonable care and, subject to remoteness rules, to pay damages for any loss or damage flowing from a breach of that obligation. If the defendant wishes to reduce its liability in a situation where the plaintiff’s own conduct contributes to the damage suffered, it is open to the defendant to make a bargain with the plaintiff to achieve that end.

In contract, the plaintiff gives consideration, often very substantial consideration, for the defendant’s promise to take reasonable care. The terms of the contract allocate responsibility for the risks of the parties’ enterprise including the risk that the damage suffered by one party may arise partly from the failure of that party to take reasonable care for the safety of that person’s property or person.” (at p 777, cols 1 and 2)

For these reasons the majority concluded that a construction applying the apportionment legislation to contract cases was “contrary to the text, history and purpose of the legislation” (p 777, col 2).

The Board notes that in their judgment the majority added:

“Perhaps the apportionment statute should be imposed on parties to a contract where damages are payable for breach of a contractual duty of care. If it should, and we express no view about it, it will have to be done by amendment to that legislation. If courts are to give effect to the will of the legislature, it is not possible to do so having regard to the terms of apportionment legislation, based on the United Kingdom legislation of 1945, and the evil that it was designed to remedy.” (at p 777, col 2)

Following the decision in Astley, legislation in the various Australian states and territories was passed to extend the ambit of contributory negligence to concurrent claims in contract (see section 101, Civil Law (Wrongs) Act 2002 (Australian Capital Territory); Schedule 1, Law Reform (Miscellaneous Provisions) Amendment Act 2000 (New South Wales); section 15, Law Reform (Miscellaneous Provisions) Act 2007 (Northern Territory); section 4, Law Reform (Contributory Negligence) Amendment Act 2001 (Queensland); section 4, Law Reform (Contributory Negligence and Apportionment of Liability) Act 2001 (South Australia); section 6, Tortfeasors and Contributory Negligence Amendment Act 2000 (Tasmania); section 4, Wrongs (Amendment) Act 2000 (Victoria); section 4, Law Reform (Contributory Negligence and Tortfeasors Contribution) Amendment Act 2003 (Western Australia).)

Astley is considered further at para 337 below. For the reasons set out there, the Board is respectfully unable to agree with the reasoning in the majority judgment in that case.

Singapore

Vesta v Butcher has been followed in Singapore. In Fong Maun Yee v Yoong Weng Ho Robert [1997] 3 LRC 138, para 51, the Singapore Court of Appeal had no hesitation in adopting it and concluding that where a defendant’s liability in contract is concurrent with an identical duty in tort, the defence of contributory negligence is available to the defendant. Similarly, in Jet Holding Ltd v Cooper Cameron (Singapore) Pte Ltd [2006] SGCA 20; [2006] 3 SLR 769 the Singapore Court of Appeal held (at para 100) that contributory negligence is available as a defence to a claim in contract where the defendant’s liability in contract was concurrent with his liability in the tort of negligence independently of the existence of any contract.

Canada

In Crown West Steel Fabricators v Capri Insurance Services Ltd (2002) DLR (4^th) 577, paras 10-21, the Court of Appeal of British Columbia noted that courts in England, Australia and New Zealand were divided on the issue of apportionment for contributory fault in contract, but followed Vesta v Butcher in allowing apportionment of damages under the relevant Act where there was concurrent liability in contract and in tort.

The text of the statute

The starting point for any consideration of the scope of the defence of contributory negligence must be the text of the statute. As was explained in Standard Chartered Bank, paras 11-12 per Lord Hoffmann, two points are important: the formulation adopted in section 1(1) of the 1945 Act and copied in section 8(1) of the 1996 Revision and the definition of “fault” in both statutes.

Section 8(1) of the 1996 Revision and section 1(1) of the 1945 Act

This point is important because the drafting of section 1(1) and section 8(1) makes it clear that what is intended is the substitution of a partial defence of contributory negligence (to the extent which is just and equitable having regard to the claimant’s share in the responsibility for the damage) for what was, at common law, a complete defence to a claim arising by reason of the contributory negligence of the claimant: “… a claim in respect of [relevant damage] shall not be defeated by reason of the fault [ie, as defined, the contributory negligence] of the person suffering the damage, but the damages recoverable in respect thereof shall be reduced [etc]” (emphasis supplied). This feature of the legislation is the basis for the decisions in Standard Chartered Bank and Co-operative Group v Pritchard: at common law, contributory negligence was not a defence to a claim in deceit or assault and battery, so section 1(1) of the 1945 Act is not applicable in relation to those torts. The legislation requires an examination of the scope of the defence of contributory negligence as a complete defence at common law, and provides that where such a defence existed there shall instead be a partial defence.

On behalf of Primeo, Mr Smith submits that whatever authoritative value Vesta v Butcher may once have had, it has now been impliedly overruled by Standard Chartered Bank and by Co-operative Group v Pritchard, a view supported by Professor James Goudkamp (“The Contributory Negligence Doctrine: Four Commercial Law Problems” [2017] Lloyd’s Maritime and Commercial Law Quarterly 213, pp 218-221). In both cases it was held that contributory negligence was not available as a defence under the 1945 Act because it would not have been available as a defence to the particular torts in issue, deceit and assault and battery respectively, before the Act was passed. The argument, as expressed by Professor Goudkamp, is that because the 1945 Act extends only to actions to which the contributory negligence doctrine extended at common law, and as (according to him) it is generally accepted that proceedings in breach of contract are not among those actions, the 1945 Act cannot apply to contractual claims. Vesta v Butcher is inconsistent with this analysis and has been overruled sub silentio. As a result, it is said, the 1945 Act does not apply to category 3 cases any more than it applies to category 1 and category 2 cases.

In Standard Chartered Bank the issue was whether a defence of contributory negligence was available to a claim in the tort of deceit. Lord Hoffmann, with whom the other members of the Appellate Committee agreed, explained that conduct of a plaintiff could not be “fault” within the meaning of the 1945 Act unless it gave rise to a defence of contributory negligence at common law. This, he considered, was in accordance with the purpose of the statute which was to relieve plaintiffs whose actions would previously have failed, rather than to reduce the damages which previously would have been awarded against defendants. The question was, therefore, whether the plaintiff’s conduct would have been a defence at common law to its claim for deceit. In the case of fraudulent misrepresentation there was at common law no defence of contributory negligence and therefore no apportionment under the 1945 Act was possible.

In Co-operative Group v Pritchard the plaintiff had been assaulted by an employee of the defendant, who alleged that the employee had been provoked by the plaintiff. The Court of Appeal held that the 1945 Act did not apply to a claim in assault and battery because prior to the 1945 Act there would have been no defence of contributory negligence available at common law in respect of such an action.

On behalf of Primeo, Mr Smith submits that it is impossible to reconcile the reasoning in these decisions with Vesta v Butcher: if a contributory negligence defence is only available where contributory negligence would have been available as a defence at common law, then it cannot be a defence to a claim in breach of contract. The submission, however, misses the point. The relevant conduct in Standard Chartered Bank was deceit, an intentional tort, which never gave rise to a defence of contributory negligence prior to the 1945 Act. Similarly, so far as Co-operative Group v Pritchard is concerned, the claim was in assault and battery, intentional torts, which also never gave rise to a defence of contributory negligence prior to the 1945 Act. In the same way, one step in Mr Smith’s reasoning is that a claim in breach of contract would not have given rise to a defence of contributory negligence prior to the 1945 Act. However, in the present case the Board is addressing a situation in which there is said to be concurrent liability in contract and in tort and where, employing the formulation of the New Zealand Court of Appeal in Rowe v Turner, it is said that negligence is an essential ingredient of both causes of action, ie a Vesta v Butcher category 3 claim. Neither Standard Chartered Bank nor Co-operative Group v Pritchard was concerned with a comparable situation. In our view the decisions and reasoning in Standard Chartered Bank and in Pritchard are not inconsistent with Vesta v Butcher.

The relevant question for consideration is, therefore, whether prior to the 1945 Act contributory negligence was available as a defence where there existed concurrent claims in contract and tort and where negligence was an essential ingredient of both causes of action. The Board will turn to this question after consideration of the definition of “fault”, to see whether that excludes the operation of the 1996 Revision (or the 1945 Act) in a category 3 case.

“Fault”

The definitions of “fault” in section 4 of the 1945 Act and in the 1996 Revision each have two limbs. (On the 1945 Act, see Glanville Williams, Joint Torts and Contributory Negligence (1951) p 318; Reeves v Commissioner of Police of the Metropolis [2000] 1 AC 360, 382C-G, per Lord Hope; Standard Chartered Bank, per Lord Hoffmann at para 11). The first limb of the definition in the 1996 Revision (“an act creating a liability in tort”) refers to the defendant’s fault, while the second limb (“an act … which, prior to the operation of this Law, would have given rise to the defence of contributory negligence”) refers to the claimant’s fault. In this the 1996 Revision resembles the differently worded definition of “fault” in the 1945 Act where the first limb (“negligence, breach of statutory duty or other act or omission, which gives rise to a liability in tort”) refers to the defendant’s fault, while the second limb (“negligence, breach of statutory duty or other act or omission which … would, apart from this Act, give rise to the defence of contributory negligence”) refers to the claimant’s fault.

When one is addressing under the 1996 Revision the question of the defendant’s fault (ie whether any person suffers damage as the result of the fault of any other person or persons), the first limb of the definition of fault (“an act creating a liability in tort”) focuses not on the cause of action on which the claim is based but on the conduct which gives rise to the cause of action. (This is also true when considering the definition of fault in the 1945 statute, but the point is even more apparent in the Cayman statute.) In his submissions Mr Gillis on behalf of the respondents was therefore correct to distinguish between the factual nature of the underlying act or omission and the resulting cause of action. The former must be capable of forming the basis of liability in tort but the latter could be a cause of action in contract or in tort. Reference is made to the law of tort for the purpose of identifying the nature of the required fault, but the resulting liability is not required to be characterised as a cause of action in tort and could include concurrent liability in contract and tort. For the defence of contributory negligence to be available, therefore, it is not necessary that the claimant sue on a cause of action in tort, but it is sufficient that the cause of action is founded on an act or omission which gives rise to liability in tort. As a result, the defence should in principle be available in category 3 cases where the claimant relies on concurrent liability in contract and in tort.

Support for this reading of “fault” as it applies to a defendant is provided by the modifications made by the 1945 Act in its application to Scotland. Section 5(a) provides:

“… the expression ‘fault’ means wrongful act, breach of statutory duty or negligent act or omission which gives rise to liability in damages, or would apart from this Act, give rise to the defence of contributory negligence; …” (emphasis supplied)

This should be contrasted with the words “negligence, breach of statutory duty or other act or omission which gives rise to a liability in tort” (emphasis supplied) in section 4 of the 1945 Act, which applies to England and Wales. It seems most improbable that the expressions “liability in damages” and “liability in tort” were intended to lead to different substantive results in the different jurisdictions. As a result, therefore, the definition in section 5 reinforces the view that what is significant here is not the characterisation of the cause of action relied upon but the nature of the underlying conduct, which must be capable of giving rise to a liability in tort.

In the course of his submissions, Mr Smith on behalf of Primeo submitted that the substitution in the definition of “fault” in the 1977 Revision of the words “an act creating a liability in tort” for the words “negligence, breach of statutory duty or other act or omission which gives rise to liability in tort” which appeared in the 1964 Law, as copied from the 1945 Act, demonstrates that the 1964 Law was only ever intended to apply to claims in tort and not to claims in contract where there exists a concurrent tortious liability. On the contrary, however, the Board considers that the words “an act creating a liability in tort” relate even more clearly to the underlying conduct as opposed to a cause of action. They refer to any act which creates a liability in tort and are apt to apply even if that act also creates a liability in contract. Neither the 1964 Law nor the 1977 or 1996 Revisions can be read as referring to an act creating a liability solely in tort.

In the same way, the second limb of the definition of fault in section 2 of the 1996 Revision, addressing claimant’s fault (“an act … which, prior to the operation of this Law, would have given rise to the defence of contributory negligence”) directs attention to an act which would at common law have given rise to the defence of contributory negligence. Once again, the focus is on the conduct in question. The same is true of the second limb of the definition of fault in section 4 of the UK statute (“negligence, breach of statutory duty or other act or omission which … would, apart from this Act, give rise to the defence of contributory negligence”), since it is clear that, in addition to the words “other act or omission”, “negligence” is a word applicable to both claimant and defendant and therefore that it refers to a standard of conduct, not a cause of action. This makes sense, in that the approaches to what constitutes fault on the part of the claimant and the defendant mirror each other for the purpose of the comparison exercise required by section 8(1) of the 1996 Revision (and section 1(1) of the 1945 Act) so that one is in that way comparing like with like: in both cases it is the conduct of the claimant and the conduct of the defendant which has to be taken into account, judged according to objective standards laid down by the substantive law applicable to the claimant and the defendant respectively. When the statutory provisions are approached in this way, it becomes apparent that, contrary to the submission of Mr Smith, there is nothing in the definition of “fault” which has the effect of preventing contributory negligence from being available as a defence where there are concurrent claims in contract and tort, where that was the case at common law before the legislation. What matters is whether the claimant’s conduct is the sort of conduct which before the legislation was enacted would have given rise to a defence of contributory negligence.

This approach to the legislation accords with that of Prichard J in Rowe v Turner Hopkins & Partners (at pp 555-556) with regard to the corresponding provisions in the New Zealand Contributory Negligence Act 1947 (sections 2 and 3(1)), in the passage cited at para 309 above, which the Board endorses.

Contributory negligence as a defence to concurrent claims in contract and tort prior to 1945

With respect to the majority of the High Court in Astley, they were wrong to say that the common law did not recognise a defence of contributory negligence to a category 3 claim framed in contract, ie where the duty of care sued upon was the same in contract and in tort. In fact, as adverted to by O’Connor LJ in Vesta v Butcher and by Beldam LJ in Barclays Bank v Fairclough Building, before the 1945 Act the common law recognised contributory negligence as a complete defence in cases involving concurrent duties of care in tort and in contract. This is clear from text-books and practitioner works in the late 19th and early 20th centuries and from a range of authorities. In the 1940s this feature of the law was well understood by practitioners and judges, as well as by the Law Revision Committee whose recommendation to change the law was adopted by Parliament in the form of the 1945 Act. It was clearly understood by Parliament, whose intention accordingly was that section 1(1) of that Act would apply in cases of concurrent duties in contract and tort, ie in category 3 cases. Hobhouse J and the Court of Appeal did not err in Vesta v Butcher.

Prior to 1945 there were well recognised categories of case in which concurrent duties of care arose in tort and contract. The leading practitioner work on the law of negligence, Beven on Negligence, 4^th ed (1928), dealt with cases in tort and in contract, using negligence as the common unifying category. According to the analysis it set out, a duty of care could arise from general circumstances in society or from special circumstances associated with free choice and from the terms of various contracts (p 15). These were examined under various headings in vol II (“Special Relations Arising Out of Contract”) such as those in Book V (“Bailments”), including common carriers (chapter 2) and common carriers by land (chapter 3, covering railway cases), and in Book VI (“Skilled Labour”) in relation to particular forms of relationship (including master and servant, doctor and patient, lawyer and client and so on). Vol I, chapter IV, entitled “Contributory Negligence”, treated that topic as applicable across all categories of case in tort and in contract where a duty of care existed; contributory negligence was described as “a principle running through the whole law of negligence” (ie in negligence cases whether founded on contract or tort). In line with this, it was stated (p 1076) that “contributory negligence is a defence in the case of common carriers as in other relations” and it was affirmed (at p 1191) that the general legal principles in relation to contributory negligence were applicable to common carriers by land, including rail travel where a duty of care arose in contract as well as in tort.

Charlesworth, in the first edition of his work The Law of Negligence (1938), also noted that in the class of case involving negligence by certain professionals and skilled persons, such as doctors, a concurrent duty of care could arise in both tort and contract: p 8 and chapter 19; and treated the defence of contributory negligence as applicable in any case of negligence, whether framed in contract or tort: chapter 22. Pollock’s Law of Torts 14^th ed (1939) by P Landon, referred to concurrent duties of care in contract and tort in certain cases (pp 348-349) and again treated the defence of contributory negligence as so applicable (pp 366-380). Salmond on the Law of Torts 10^th ed (1945) by W.T.S. Stallybrass, 1945, which on this point was the same as the 9^th edition of 1936, identified cases of concurrent duties of care in contract and tort in negligence (pp 8-9) and did not suggest that the defence of contributory negligence was unavailable in relation to the former.

To similar effect, a lengthy American treatment of the subject which reviewed English and US cases explained that the defence applied to concurrent claims in contract and tort: Charles Fisk Beach Jr, A Treatise on the Law of Contributory Negligence (1885), p 6. The ability of an employee to sue for negligence in both contract and tort was recognised in W. Roberts and G. Wallace, The Common Law and Statutory Duty and Liability of Employers 4^th ed (1908), p 177 (it being noted that by reason of a more generous costs regime applicable to tort claims at the time, such claims tended to be analysed as claims in tort); and contributory negligence was treated as a defence however the claim was framed. Roberts and Wallace stated (p 159), “[t]he relation of master and servant being a matter of contract, the extent of the master’s duties toward the servant must depend on that contract”, with reference to M’Laughlan v Dunlop, 20 SLR 271, in which Lord Young observed (p 275) that at common law the relationship “is all a matter of contract”. Glanville Williams, the leading commentator on the 1945 Act, referred to a range of cases, including railway cases, to conclude that contributory negligence was a defence to concurrent claims in contract and hence that the 1945 Act covered such claims: Joint Torts and Contributory Negligence, pp 214-222 and 328-332.

There are many authorities which support this view. It is sufficient to focus on two areas to make the point, the employment relationship (the relationship of master and servant as it was called in the nineteenth century) and the railway cases. In the first, an employer owed his employee a duty of care in various respects under contract and tort and pursuant to statutory duties which came to be imposed, but employee claims could be defeated on the grounds of contributory negligence. In the second, a railway company owed a duty of care to its passengers in contract and in tort to transport them safely, but a claim by a passenger for breach of this duty could again be defeated on the grounds of contributory negligence.

In Thomas v Quartermaine (1887) 18 QBD 685 the Court of Appeal reviewed the operation of the defence of contributory negligence in a master and servant case involving an injury suffered by the servant/employee. At trial the employee was acquitted of contributory negligence, but by a majority the Court of Appeal held that the master/employer could still rely on a defence of volenti non fit injuria on the particular facts, as the employee had been fully aware of the risk involved when acting. The court recognised that contributory negligence was a general defence in master and servant cases, without seeking to categorise an employee’s claim as being framed either in contract or in tort. As Bowen LJ observed (p 694), in such a case the employee would have to prove that the defendant “had been guilty of some negligence, that is to say, of some breach of duty towards the plaintiff himself” (ie without needing to specify whether it was a duty arising in tort or in contract) and also that “the defendant’s negligence had been the proximate cause of such an accident”, and contributory negligence was relevant only to the latter requirement: “[c]ontributory negligence in a plaintiff only means that he himself has contributed to the accident in such a sense as to render the defendant’s breach of duty no longer its proximate cause” (see also p 697). Clearly, on this analysis, contributory negligence would be a defence whether the negligence claim was framed in contract or in tort. At pp 698-699 Bowen LJ said: “The law is full of instances where duties assume a double aspect and may be viewed concurrently as arising by implication out of a contract, or as created by some wider principle of law which happens to take effect and to receive apt illustration in the particular instance of some particular contract. It is in most cases a barren and metaphysical inquiry to discuss whether such duties are best treated as arising by implication from the contract or from the general law outside …”. Albeit this was said in relation to the interaction between the defences of contributory negligence and volenti non fit injuria, the observation again indicates an awareness that a relevant duty of care could arise concurrently in contract and tort without the need to assign it definitively to one category or the other.

The defence of contributory negligence was often applied without analysis of which category, contract or tort, might be applicable: see, eg, the account by A.L. Smith LJ in Weblin v Ballard (1886) 17 QBD 122, 124, of the position at common law prior to the Employers’ Liability Act 1880 (43 & 44 Vict c 42). As Lord Radcliffe later explained in Lister v Romford Ice and Cold Storage Co Ltd [1957] AC 555, 587, in a manner reminiscent of Bowen LJ’s analysis, “[t]he existence of the duty arising out of the relationship between employer and employed was recognised by the law without the institution of an analytical inquiry whether the duty was in essence contractual or tortious. What mattered was that the duty was there.” In Wilson v Merry & Cunningham (1868) LR 1 HL Sc 326, 332, Lord Cairns LC explained that “[t]he master is not, and cannot be, liable to his servant unless there be negligence on the part of the master in that in which he, the master, has contracted or undertaken with his servant to do”. This statement, which was the foundation for the common law common employment defence, has been described as “[p]erhaps the most important – and certainly the most influential – nineteenth-century exposition of employers’ common law duties”: Paul Mitchell,A History of Tort Law 1900-1950 (2015), 229.

This analysis of the position was current in leading authorities on the employment relationship in the period leading up to the 1945 Act. An employer’s obligation to exercise reasonable care was treated as a matter of contract in Fanton v Denville [1932] 2 KB 309, 314 (Scrutton LJ) and 335 (Slesser LJ), a case in which a defence of contributory negligence had been in issue, though not made out on the facts. It was so treated again in the leading case of Wilsons & Clyde Coal Co Ltd v English [1938] AC 57 (see eg the discussion of contract law at p 88 per Lord Maugham), in which Fanton v Denville was considered and disapproved. English was another case in which a defence of contributory negligence had been in issue, though not made out on the facts. The House of Lords did not suggest that the defence was not available in such cases.

Similarly, in railway cases the point was made that a claim for breach of a duty of care by a railway company could be framed in contract or in tort: see eg Foulkes v Metropolitan District Railway (1880) 5 CPD 157, 164 (per Baggallay LJ) and p 170 (per Thesiger LJ) and Kelly v Metropolitan Railway Co [1895] 1 Q.B. 944, 946 (Lord Esher MR); and in the authorities the defence was frequently applied without analysis of which category might be applicable. It did not matter, because the defence was applicable in both cases.

In the early 20th century editions of Bullen & Leake’s Precedents of Pleadings it was stated that “[a] carrier is in many cases liable to be sued either in contract or in tort at the option of the plaintiff”, with it being “generally sufficient for the pleader merely to allege the material facts, disregarding all questions of the precise form of action”: 7^th ed (1915), p 105; 8^th ed (1924), pp 144-145; and 9^th ed (1935), pp 136-137. In the section on “Statements of Claim in Actions of Contract” there is reference to claims for damages for personal injuries sustained by a passenger in a collision caused by a railway company's negligence: pp 114, 155 and 147, respectively; this type of claim does not have its own precedent, but the reader is referred to one in the section on “Statements of Claim in Actions for Torts”.

Use of the word “negligence” in a claim did not involve the election of a cause of action in tort rather than in contract, because in both cases the obligation was to carry without negligence. In most of the cases the passenger’s claim is simply discussed as a claim for negligence. However, Gee v Metropolitan Railway Co (1873) LR 8 QB 161 is an example of a claim brought in contract (see p 164) in which the questions of negligence by the railway company and contributory negligence by the passenger were both held to have been properly left to the jury (see Cockburn J at 165-166, Blackburn J at 166, Kelly CB at 171, Keating J at 174, Brett J at 176 and Clearsby B at 177-178). Bridges v North London Railway Co (1874-75) LR 7 HL 213 is another example in which the liability of the railway company was said to be based simply on negligence for the most part, in relation to which contributory negligence was recognised to be a defence, where one of the judges (Brett J) expressly held that the railway company’s liability was in contract and that contributory negligence would be a defence to such a claim (pp 231-232).

Bullen & Leake stated that contributory negligence could be pleaded as a defence to an action for certain breaches of contract. In the section on “Defences, etc, in Actions of Contract”, there was included the “Defence of Contributory Negligence to a Claim in respect of Personal Injuries”: this defence did not have its own precedent, the reader being referred to one in the section on “Defences, etc, in Actions for Tort” (pp 548, 651 and 650, respectively). In the editions of Chitty on Contracts in the first part of the 20th century, after noting that a railway company’s contract was to carry without negligence, it was stated that “[a] plaintiff is precluded from recovering by his contributory negligence”: Chitty on Contracts 18^th ed (1930), pp 549-550 and 553; 19^th ed (1937), pp 726 and 730; earlier editions are to the same effect (16^th ed (1912), p 523; 17^th ed (1921), p 558).

Legislative history of the 1945 Act

This view of the contemporary legal position and contemporary understanding of the law in the mid-1940s is supported by the history of the legislation which became the 1945 Act. The 1945 Act was intended to reform the law in the light of the Law Revision Committee’s Eighth Report (Contributory Negligence) Cmnd 6032 (1939). The background is explained in Mitchell,A History of Tort Law 1900-1950, ch 13 (Contributory Negligence) and Jenny Steele, “Law Reform (Contributory Negligence) Act 1945: Collisions of a Different Sort”, in A. Arvind and J. Steele (eds) Tort Law and the Legislature: Common Law, Statute and the Dynamics of Legal Change (2013).

The Bill as introduced into Parliament reflected the primary focus of the Law Revision Committee’s report. However, as Professor Steele explains, changes were made as the Bill went through Parliament to try to ensure that the new law would apply in all standard cases where the defence of contributory negligence was available. Viscount Simon LC, who introduced the Bill in the House of Lords, is recorded as having pointed out that “many claims for damages for personal injuries were laid both in tort and as a breach of contract to carry safely, and a plea of contributory negligence could be raised which, if successful, would defeat the plaintiff”, so there is no doubt that Parliament appreciated that the defence of contributory negligence applied in cases of concurrent duties. Viscount Simon made it clear that the intention was that the new law would apply in all cases where the defence had previously been available and for this reason an earlier proviso that the statutory provision “shall not apply to a claim under a contract” was deleted (Steele, ibid, pp 182-3).

Professor Mitchell explains the background to changes in what became section 2 of the 1945 Act as it proceeded through Parliament. In the course of drafting the Act concerns had been raised about its effect on parallel claims brought by workers against their employers for negligence under the Workers Compensation Act 1925, which enabled claims for compensation (albeit at a lower level than under the common law) in relation to which contributory negligence was not a defence, provided no compensation was awarded in any common law claim. The concern was that if contributory negligence were changed from being a complete defence to a partial defence, a workman might obtain some damages at common law, possibly at a rate lower than in a claim under the 1925 Act, with the result that he would lose the benefit of a claim under that Act. Therefore, the original version of section 2(1) in the Bill as introduced in Parliament provided that the 1945 Act would not apply in actions by workmen against employers: see Mitchell, op. cit., pp 323-324; and Viscount Simon, HL Deb 23 January 1945, vol 134, col 636. Steele explains (p 179) that there was a political backlash, with the result that Viscount Simon introduced an amendment which, he explained (HL Deb 6 June 1945, vol 136, col 404), would ensure that if a workman found that he was going to get only a very small sum at common law as a result of application of the 1945 Act he would be entitled at that stage to choose instead to make his claim under the 1925 Act. Section 2 of the 1945 Act was enacted in this form. It was repealed, along with the entirety of the 1925 Act, by the National Insurance (Industrial Injuries) Act 1946, which was commenced in 1948. Since section 2 was not on the statute book for very long, this episode has sometimes been overlooked. What is significant about it is that each twist in the drafting of this part of the 1945 Act was predicated on the understanding that contributory negligence was, at common law, a defence to a claim brought by a worker against his employer for negligence, whether framed in tort or in contract. There was no suggestion that a worker had the option to avoid the operation of the defence of contributory negligence by the simple expedient of suing in contract.

In the Board’s view, the respondents are correct in their submission that the removal of the words “Provided that this subsection shall not apply to any claim in contract” in clause 1 and the changes in the drafting of what became section 2 show that the 1945 Act was enacted with concurrent claims in contract and tort in mind and with the intention of making the defence of contributory negligence available in circumstances where a concurrent claim existed.

Certainty in the law

For the reasons set out above, the Board has come to the firm conclusion that Vesta v Butcher was correctly decided and has a sound basis in principle. A further consideration to which the Board attaches considerable importance is the fact that Vesta v Butcher has been accepted as settled law for some 35 years, both in England and Wales and in a number of other common law jurisdictions. Throughout this period contracting parties, both in the area of the provision of professional services and more generally, have negotiated and contracted on the basis that where there may arise concurrent claims in contract and in the tort of negligence, contributory negligence will be available as a defence. In this regard the Board draws attention to the striking fact that when in 1989 the Law Commission of England and Wales conducted a public consultation on the availability of contributory negligence as a defence in contract (Law Commission Working Paper No 114, Contributory Negligence as a Defence in Contract) not a single respondent thought that apportionment should be excluded in Category 3 cases. It was acknowledged that it would be undesirable to have contributory negligence as a defence in one cause of action and not in another arising out of the same facts (Law Commission, Contributory Negligence as a Defence in Contract, Law Com No 219, 1993, para 3.42). See also the report of the Scottish Law Commission on Civil Liability Contribution (Scot Law Com No 115, 1988) which recommended that where the defender’s liability for breach of a contractual duty of care is the same as his liability in delict for negligence, the plea of contributory negligence should be available as a defence whether the action is framed in delict or in contract (para 4.15, Recommendation 20). The Commission observed that although the 1945 Act may be open to this interpretation, it considered that the matter should be put beyond doubt.

The claim against HSBC and whether a defence of contributory negligence is available

The next question is whether the defence of contributory negligence is available to HSBC, as custodian, on the basis that Primeo’s claim for contractual damages falls in Vesta v Butcher category 3 or for some other reason.

The trial judge, Jones J, drew a distinction between the administrator and the custodian. The case against the administrator, the Bank of Bermuda, was relatively straightforward: the claim in respect of the NAV calculations was based upon an implied term of the contract that it would exercise reasonable skill and care. This contractual duty was co-extensive with the tortious duty of care which arose from the Bank of Bermuda’s position, for it was to be regarded as holding the office of administrator under Primeo’s articles of association, and here too it was required to act with reasonable skill and care.

The position of HSBC as custodian was different. Clause 16B of the 1996 Custodian Agreement conferred on the custodian the power to appoint suitable “agents, sub-custodians and delegates as it might think fit” to perform its duties. It was also agreed:

“… The Custodian [HSBC] will use due care and diligence in the appointment of suitable sub-custodians and must be satisfied for the duration of the sub-custody agreements as to the ongoing suitability of the sub-custodians to provide custodial services to the Company [Primeo] …. [and] will require the sub-custodian to implement the most effective safeguards available under the laws and commercial practices of the sub-custodian’s jurisdictions in order to ensure the most effective protection of the Company’s assets.”

These duties were referred to by the trial judge as, respectively, “the appointment and ongoing suitability” and “the most effective safeguards” duties. Further, in the words of Clause 16B:

“Subject thereto the Company [Primeo] has agreed to indemnify the Custodian [HSBC] from all liabilities of whatsoever nature which may be [in]curred by it in performing its obligations under the Custody agreement other than those liabilities resulting from fraud, negligence or wilful breach of duty on the part of the Custodian or any agent appointed by it. …”.

Clause 16E of the 1996 Custodian Agreement is also important and provides, so far as relevant:

“The Custodian shall not, in the absence of negligence or wilful breach of duty on the part of the Custodian or any agent, delegate or sub-custodian, be liable to the Company ... for any act or omission … in the course of or in connection with the services rendered by it hereunder or for any loss or damage which the Company may sustain or suffer as a result or in the discharge by the Custodian of its duties hereunder or pursuant thereto.... The Company agrees to indemnify the Custodian from and against any and all liabilities, obligations, losses, damages, penalties, actions, judgments, suits, costs, expenses or disbursements of any kind or nature whatsoever (other than those resulting from the negligence or wilful breach of duty on the part of the Custodian or any agent appointed by it) which may be imposed on, incurred by or asserted against the Custodian in performing its obligations or duties hereunder. ”

It can be seen now that clause 16B contains three elements which have a bearing on this aspect of the appeal. First, it conferred on the custodian a power to appoint agents, sub-custodians and the like to perform the duties for which it was responsible. Secondly, it imposed on the custodian a duty to take reasonable care and to exercise due diligence when appointing, among others, sub-custodians. Thirdly, it imposed on the custodian an obligation to require any sub-custodian to implement the most effective safeguards available under the laws and practices of the appropriate jurisdiction.

It was the third of these elements which proved particularly important. HSBC, as custodian, owed a continuing duty to Primeo to satisfy itself about the ongoing suitability of BLMIS as sub-custodian to provide custodial services to Primeo, and further and most relevantly for present purposes, it owed an obligation to Primeo in relation to the implementation of the most effective safeguards available under the laws and commercial practices of New York in order to ensure the protection of Primeo’s assets.

The trial judge also explained, at para 193, that in the circumstances of this case, where the sub-custodial service provided to HSBC was part of an indispensable package of services provided to Primeo, it was not within the power of HSBC, as a practical commercial matter, to require BLMIS to do anything without the agreement of its client, Primeo. Nevertheless, the judge was satisfied that there was an obligation on HSBC to recommend to Primeo the relevant safeguards which it ought to implement, and a failure to make the recommendation would constitute a breach.

The judge summarised the position at paras 324 and 325 in the following way: on its true construction, clause 16B created ongoing duties which required periodic performance. HSBC was in breach of the clause when it failed in August 2002, June 2003, March 2004, March/April 2005 and February 2007 to give any consideration or make any recommendations to Primeo about safeguards which were readily available and, if implemented, would have been effective to safeguard Primeo’s assets.

Primeo accepted that it was an implied term of the contract that in performing these duties and obligations HSBC was required to exercise reasonable care and skill. But this was not the same as a duty to safeguard the assets in a manner to be expected of a reasonably competent global custodian. Primeo maintained that, for these reasons, a plea of contributory negligence was not available to HSBC in respect of the claim against it under Clause 16B. The judge agreed.

The Court of Appeal came to a different conclusion. It explained (at paras 484-497) that in determining the nature of a defendant’s contractual liability it was necessary to focus on the language of the contract as a whole. It also recognised that in order to decide whether the liability in contract fell within category 3 of Vesta, the liability in contract must be the same as the liability in the tort of negligence independently of the existence of any contract. These are propositions with which the Board would agree.

Ultimately the Court of Appeal considered the question turned on the construction of clause 16B and whether the most effective safeguards duty was co-extensive with a duty in tort. In its view, the effect of the final sentences of each of clauses 16B and 16E of the Custodian Agreement was that HSBC was entitled to be indemnified by Primeo against any liability incurred by HSBC except where HSBC had acted negligently, fraudulently or in wilful breach of duty. It followed that if HSBC had committed a non-negligent, non-deliberate breach of duty, any claim by Primeo against HSBC would be defeated by circuity of action since HSBC would be entitled to enforce the indemnity. As the Court of Appeal put it, HSBC’s liability would only arise where it had failed to take appropriate care.

The Court of Appeal therefore decided that the overall effect of clause 16B in the light of the exonerating provisions in it and in clause 16E was that HSBC was not liable unless it was negligent. Its liability arose from a contractual obligation which in substance was to exercise reasonable skill and care in connection with the most effective safeguards, and so fell in category 3.

The Board recognises the force of the arguments accepted by and the reasoning of the Court of Appeal but has reached the conclusion that the judge approached the matter correctly, and that the Court of Appeal was wrong to allow the appeal on this issue. That is so for the following reasons, which reflect those developed and relied upon in submissions made to the Board on behalf of Primeo.

First, there was never a contract between Primeo and HSBC, as custodian, under which HSBC was required to perform professional custodial services subject to a relevant express or implied obligation to exercise reasonable skill and care, and where there was a concurrent duty of care in tort to exercise reasonable care in the provision of those services. The relevant duty imposed on and undertaken by HSBC under the Custodian Agreement was a duty to do something specific, that is to say, at the least, to require the sub-custodian to put in place the most effective safeguards in order to ensure the most effective protection of the assets. In the context of this case, that meant making the appropriate recommendations to Primeo, and that is what it failed to do. No basis has been shown for categorising this as anything other than a Vesta category 1 case, or perhaps a category 2 case.

Secondly, the nature of the most effective safeguards duty was not transformed into a duty to take reasonable care in relation to the provision of a service because of the existence of the indemnity and exoneration provisions. This is to confuse the nature of the duty and the issue of breach of the duty with the question of the ultimate liability of HSBC as custodian. Put another way, the exoneration and indemnity provisions do not change the scope and nature of the duty. Instead, they deal with the position once a breach of duty has been established and operate to limit the circumstances in which damages may be payable for that breach.

Thirdly and importantly, there was an overarching failure by the respondents to identify what the concurrent duty in tort was in this case, or what the foundation for it was said to be. This is not something with which the Court of Appeal ever properly grappled. No basis was ever established for the imposition of a duty of care in tort which was or would have been concurrent with the relevant contractual duty identified by the trial judge or the Court of Appeal.

In all these circumstances, the Board is satisfied that the trial judge was entitled and right to find that the plea of contributory negligence was not available to HSBC in respect of the claim against it under clause 16B, and that Primeo’s appeal on this issue must be allowed.

The Court of Appeal’s approach to the contributory negligence reduction in the claim against the administrator

The judge held that Primeo’s damages against the Bank of Bermuda under the administration agreement fell to be reduced by 75% because it was, to a large extent, the author of its own misfortune.

On appeal to the Court of Appeal, Primeo contended that the judge’s decision to make a reduction of this size fell outside the broad range of outcomes within which reasonable disagreement was possible. That being so and the judge having fallen into error, the Court of Appeal should undertake a fresh assessment.

The Bank of Bermuda responded that the judge was directed to the correct principles and that he applied them to the relevant facts. In those circumstances, he was entitled to conclude that Primeo was indeed to a substantial degree the author of its own misfortune and that a deduction of 75% was entirely appropriate. There was no basis for interfering with that assessment.

The Court of Appeal agreed with Primeo and its reasons for doing so were concise and clear. In summary, the Bank of Bermuda was a professional administrator and it was performing a service which included the specific task of monitoring BLMIS. It had responsibility for calculating and issuing the NAVs and it carried out that task knowing that nothing had been done to verify the existence of the assets which formed the basis of that calculation. Nevertheless, the Bank of Bermuda did not make Primeo aware of the concerns to which these matters had given rise and which had been expressed internally within the Bank of Bermuda and HSBC about the concentration of functions and lack of transparency in BLMIS, and that there was a risk that the assets did not exist. To the contrary, Mr Fielding provided reassurance to Primeo at meetings of its Board in 2003 and 2004.

The Court of Appeal concluded that in light of all of these matters, attributing 75% of the fault to Primeo was well outside the broad range of outcomes within which reasonable disagreement was possible and was clearly wrong. It therefore fell to the Court of Appeal to make its own assessment. It had in mind that the factors which it was required to consider were incapable of precise measurement and that Primeo was keen to invest only in BLMIS. It concluded that the Bank of Bermuda’s conduct played at least an equal role to that of Primeo in causing the damage and that it was at least equally blameworthy. This was substantially different from the view that Primeo was 75% responsible and also warranted the conclusion that the judge had gone wrong. The damages awarded against the Bank of Bermuda should therefore be reduced by 50%, not 75%.

The Bank of Bermuda now contends before the Board that the Court of Appeal had no basis for interfering with the judge’s assessment. Mr Gillis submits that the judge properly took into account a range of relevant factors including the following:

Mr Gillis also submits that the reasons given by the Court of Appeal for finding the judge had fallen into error were themselves wrong. In particular, the fact that the Bank of Bermuda was providing a professional service was no bar to a finding that Primeo was primarily responsible for the damage it had suffered. Secondly, the Bank of Bermuda was not given the specific task of monitoring BLMIS and had no formal role in the identification and prevention of fraudulent activity. Thirdly, the Bank of Bermuda had no contractual obligation to provide any risk management advice to Primeo; there was an equivalence of knowledge as between them; and the Court of Appeal also wrongly took into account matters which did not amount to gross negligence.

What is more, Mr Gillis continues, the Court of Appeal failed to provide any sufficient reasons or explanation for its own view that Primeo and the Bank of Bermuda were equally responsible for the damage Primeo suffered.

The Board has no difficulty accepting that the judge took into account many relevant factors but there was one particularly important matter to which, at least in this context, he failed to attach any proper weight at all, namely that the Bank of Bermuda was a professional service provider and that it was engaged to carry out and provide fund administration services to Primeo. As Mr Smith submits and the Board accepts, an important part of the Bank of Bermuda’s professional obligation was to produce and issue a NAV for the fund. Primeo, on the other hand, had no parallel responsibility to verify that valuation. Further, the Bank of Bermuda was required to take reasonable steps to satisfy itself that the published NAV was accurate, and to do this it needed to take appropriate steps to confirm that the information it had received was complete and accurate, and that the assets did exist. That was something it failed to do.

This was a critical point, and it is one which the judge failed properly to take into account. In these circumstances the Court of Appeal was entitled to revisit the issue and to find that the Bank of Bermuda played at least an equal role to that of Primeo in causing the damage, and that it was at least equally blameworthy. Crucially, it was grossly negligent from April 2005 in issuing the NAV, knowing that there had been no independent verification of the existence of the assets.

The Board is also satisfied that the Court of Appeal was therefore entitled to find that the deduction of 75% fell outside the range of reasonable outcomes, and that an appropriate reduction was 50%. For all of these reasons, the Board rejects this aspect of the respondents’ challenge to the judgment and order of the Court of Appeal.

The Board’s conclusions can be summarised as follows.

The claims against HSBC

Primeo suffered recoverable loss: paras 57-66 above. The breach of BLMIS’s custodial duties was not rectified by the Herald Transfer: paras 67-70 above. There was an immediate loss when Primeo’s cash was misappropriated each time it made an investment in BLMIS: paras 71-96 above. Primeo did not suffer its loss in its capacity as a shareholder of Herald: para 97 above. However, the loss of cash by Primeo was mitigated by sums it received from BLMIS in the period between August 2002 and May 2007: para 98 above.

It is not open to Primeo to argue that HSBC assumed responsibility as custodian for the assets purportedly held by BLMIS on 7 August 2002: paras 157-169 above. The Court of Appeal erred in directing that questions of appropriation in general, the existence of a running account and whether the rule in Clayton’s Case was to apply should be referred to the judge assigned to deal with the assessment of damages. Primeo should not have been permitted to raise these issues before the Court of Appeal: paras 170-175 above.

In relation to causation, the Court of Appeal erred in allowing Primeo to advance a case of a loss of a chance: paras 176-190 above. Since it is not open to Primeo to argue causation on a loss of a chance basis, it is unnecessary for the Board to resolve a number of issues which would arise only in that eventuality: paras 191-195 and 199- 208.

It is also too late for the respondents to seek to argue that Primeo assigned to Herald all rights and remedies in relation to the Direct BLMIS Investments, pursuant to the Herald Transfer: paras 209-213 above.

The administration claim against Bank of Bermuda

Bank of Bermuda was negligent from 2002: paras 110-120 above. It was grossly negligent from April or early May 2005: paras 121-138 above. It was grossly negligentfrom February 2006: paras 139-143 above.

In relation to causation, it is not open to Primeo to argue a case based on the loss of a chance: paras 196-198 above.

Limitation

Any fault-based causes of action which arose prior to 20 February 2007 are time-barred. Primeo cannot rely upon section 37(2) of the Limitation Act: paras 217-236 above. On the other hand, the strict liability claim against HSBC is not time-barred, since Primeo can rely upon section 37(1)(b): paras 237-281 above.

Contributory Negligence

The defence of contributory negligence is in principle available where a claim is based on the breach of a contractual duty of care which is concurrent with a duty in tort: paras 288-353 above. It therefore applies to Primeo’s claim against Bank of Bermuda. However, it does not apply to Primeo’s claim against HSBC, first because HSBC’s duty was not one of reasonable care, but was a specific duty to require the sub-custodian to put in place the most effective safeguards, and secondly because it has not been established that HSBC was subject to a concurrent duty in tort: paras 354-371 above. The Court of Appeal’s decision that the appropriate reduction for contributory negligence in respect of the claim against Bank of Bermuda was 50% is upheld: paras 372-382 above.

The appeal

These conclusions entail the rejection of Primeo’s grounds of appeal, other than on the issues of reflective loss and HSBC’s defence of contributory negligence, on which it has succeeded. They also entail the acceptance of some, but not all, of the additional grounds on which the respondents sought to uphold the Court of Appeal’s decision. The Board will humbly advise His Majesty that the formal result is that the appeal is allowed in part.