Assume that a person (P) has a claim against another (D) for a loss suffered as a result of a breach of duty owed by D to P. Assume further that a company (Co) has a claim against D for loss suffered by Co as a result of a breach of separate duty owed by D to Co. Assume, finally, that the losses suffered by P and Co overlap, wholly or in part, and that P is a shareholder or creditor of Co. Before the decision of the Supreme Court in Sevilleja v Marex Financial Ltd ( UKSC 31), it was arguable that British law contained a rule which excluded P from recovering from D the amount of the overlapping loss. This was a strong rule. It did not state that Co’s claim in relation to the overlapping loss merely had priority. P was excluded from recovery even if the company did nothing to recover the loss (subject to one exception).
A core example of the problem was presented by the facts of Prudential Assurance Co Ltd v Newman Industries Ltd (No 2) ( Ch 204), the first case to identify the issue. D in breach of their duties as directors of Co transferred an asset of Co at an undervalue to an affiliated company which was in financial difficulty. In order to obtain shareholder approval for this transaction (as required by the Stock Exchange’s Listing Rules) D secured that Co issue a misleading circular to its shareholders, in breach of their duty to the shareholders when advising them how to exercise the rights attached to their shares. However, the hypothetical can arise whether or not D is a director of Co. Thus, the rule was applied in Johnson v Gore Wood & Co ( 2 AC 1) even though D was a corporate outsider (a firm of solicitors). This decision in fact provided the strongest support for the suggested rule.
In Prudential there was, therefore, a single scheme to transfer assets at an undervalue out of the company, but its implementation involved two separate legal wrongs. The wrong to Co caused loss to it via a reduction in its net asset value and, possibly, by a reduction of its capacity to earn profits in the future. The wrong to P as shareholder caused loss, at least hypothetically, via a drop in the value of P’s shareholding as the loss to the company fed through into the valuation of P’s shares. The point seized upon by the Court of Appeal in that case was that recovery by Co from D of the value removed from it would restore, in whole or in part, the loss suffered by P through the drop in the value of the shares.
Two points should be noted at the outset about the reasoning in Prudential. First, the relationship between the damage to Co and the value of its shares was not explored in detail, apparently because P put in no evidence that there had been a drop in value. The main goal of P in that case was to enforce Co’s rights against D in a derivative action; the assertion of P’s claims against D in its own right was a side-show. The issue was discussed in the Johnson case, but only in a rather crude way. Second, while Co’s recovery against D might redress P’s loss, in whole or in part, by increasing the value of shares now held in a company whose net asset value had been increased in comparison with the pre-recovery position, the opposite was not true. This was a one-way street. Recovery by P from D of the drop in value of its shares would not restore to Co any of the value it had lost. As we see below, the asymmetrical operation of the recovery process lies at the heart of the difficulties created by the Prudential decision. Given this, the terminology adopted in the earlier case-law to characterise P’s loss – that it was ‘reflective’ of the loss suffered by the company – does not seem inappropriate. After all, the moon reflects the rays of the sun, whether the moon is full or in another of its twelve phases. However, use of the term was deprecated in Marex, and so ‘overlapping’ will be used in this note.
The majority view
After Marex, it is clear that British law no longer contains the rule set out in the first paragraph. However, there is a distinct difference of opinion between the majority (led by Lords Reed and Hodge) and the minority (led by Lord Sales) as to the state of the current law. The minority think that no version of the suggested rule exists and that the difficulties its absence may generate can be handled effectively by other rules of law. The majority decide that an amended version of the suggested rule is still part of British law. The amendment consists of altering the rule so that it applies only where P is, and sues as, a shareholder of Co. Since in Marex P was a creditor but not a shareholder of Co, the claimant succeeded on either view. As a result, the decision reads more like a report of a high-level university seminar than anything else.
Lord Reed founded himself on the decision of the Court of Appeal in Prudential Assurance Co Ltd v Newman Industries Ltd (No 2). This was a decision about the rule in Foss v Harbottle ((1843) 2 Hare 461), with which generations of company law students used to grapple, until their attention was deflected, probably with relief, to Part 11 of the Companies Act 2006, which adopted a different approach to the issues which Foss v Harbottle sought to address. For Lord Reed, therefore, the amended rule is ‘a rule of company law, applying specifically to companies and their shareholders’ (at ), whereas for the minority there are only general rules about damages (though, perhaps, rules that are applied in a particular way in the company context). One minor quibble with Lord Reed’s characterisation of the rule in Foss v Harbottle as a rule of company law is that it is perhaps better seen as a rule about organisations, not just about companies. The rule has always been applied to, for example, trade unions, which are not companies or even corporations (see the Trade Union and Labour Relations (Consolidation) Act 1992, s 10) and never have been except for a brief period under the Industrial Relations Act 1971-1974. Indeed, Edwards v Halliwell ( 2 All ER 1064) was cited in Marex without demur as a leading case on the rule, but the relevant organisation in that case was the National Union of Vehicle Builders. No doubt, this terminological point will be ironed out in due course.
At first sight, the rule in Foss v Harbottle provides unpromising ground for addressing the issue under consideration. The rule has two limbs. The first, and more important, confines (or, rather, confined) the capacity of an individual shareholder to enforce corporate rights to a very narrow set of circumstances. Clearly, however, in our hypothetical P is not enforcing the company’s rights but P’s own rights. The enforcement of personal rights has always been treated as an exception to the first limb of the rule: see Edwards v Halliwell. The second, less important and more obscure, limb of the rule puts impediments in the way of a shareholder who wishes to complain about some impropriety in the conduct of the company’s internal affairs, usually by way of seeking an injunction to restrain the company and its officers from acting on a shareholder resolution adopted in some procedurally incorrect manner. Here, as well, personal rights have long been regarded as not triggering these impediments, though the question of when a personal right will arise in this context has never been clearly answered by the courts. See Pender v Lushington (1877) LR 6 Ch D 70. In any event, our hypothetical is not one in which P is seeking to complain about some internal impropriety of Co.
Lord Reed’s intellectual leap here is to link the rule excluding P from recovering overlapping loss from D to the rationale of the first limb of the rule in Foss v Harbottle. This is the step he perceived the Court of Appeal to have taken in the Prudential case. Whether that court did take this step is open to some doubt. Since P in that case, as noted above, did not lead any evidence about the fall in the value of its shares, the court may have thought that Prudential was seeking to recover in a personal action against D losses which had been suffered by Co, something that was clearly illegitimate. It is interesting that the Law Commission in its Consultation Paper on Shareholder Remedies in 1996 (CP 142) devoted only a single descriptive sentence to this aspect of the Prudential case. One would have expected that body to say more had it perceived the case to have adopted a rule about loss exclusion which constrained shareholders’ remedies in so important a way. The final report did not even mention the issue.
However, the more interesting question is whether the loss exclusion rule fits into the rationale for the first limb of the rule in Foss v Harbottle. Lord Reed, following Lord Bingham in Johnson, locates the rationale of the first limb in the goal of protecting the autonomy of the company (at ). This seems to me to be correct. A company is an organisation, not an ad hoc collection of unrelated individuals. Co’s right of action against D is a form of corporate property. How that property is to be exploited should surely be a matter for a body which can claim to speak on behalf of the organisation as a whole and not a matter for an individual member of it – unless there is some good reason for thinking in particular circumstances that the decision of any available representative body would be unsafe or unreliable.
The tricky bit, of course, is how one moves from this proposition to the conclusion that the autonomy rationale requires the application of the loss exclusion rule when P sues D for breach of P’s own rights against D. It is the asymmetrical operation of the recovery process which provides the impetus to take this step. The exact relationship between the company’s loss of assets and the decline in the value of its shares is debated at length in Marex, but does not really affect the point of principle. It is clearly not a one-to-one relationship. If, as the finance people tell us, the value of a share, at least in a publicly traded company, reflects the present value of the expected future stream of distributions from the company, the loss of assets will impact only indirectly on the share price and the size of the impact will vary according to the circumstances. However, in the case of a significant loss of assets, there is unlikely to be no relationship at all, and P is likely to sue D only where that impact is substantial. The rules on causation are enough to deal with cases where the impact is undetectable.
As Lord Reed sees it, rightly in my opinion, the autonomy issue arises in this way. Assuming a rule protecting D from double recovery, if the rule excluding recovery of overlapping loss in P’s action against D did not apply, then assets (or their value) which before the wrongdoing were held by the company might end up in the hands of the suing shareholders after the litigation. This would obviously be bad for creditors of Co, since assets would have been moved from a defendant against whom they can assert their claims into the hands of P against whom this is not normally possible. However, whether creditors have a legitimate objection to this move depends upon whether Co has distributable profits which cover the loss of assets. If it does, creditors, it might be said, have no ground of complaint if the result of the litigation is in functional terms a distribution of Co’s assets to the litigating shareholders. Certainly, the story is different if Co does not have distributable profits, but it can be seen that creditor protection provides an incomplete rationale for the rule excluding across the board recovery of overlapping loss.
What about shareholders? The autonomy argument here, again assuming a rule against double recovery, is that failure to exclude overlapping loss in P’s litigation takes out of the hands of a representative body of the company (the board or the shareholders acting together) the decision as to how Co’s assets (Co’s claim in this case) are to be deployed and places that decision in the hands of individual shareholders. If P can recover the overlapping loss from D before Co gets to D, the decision over the use of that asset has moved from Co to P. P’s recovery undermines the autonomy of the company in two ways. First, the representative body no longer has full control over Co’s assets (typically vested in the board by the articles). Second, there is an undermining of the principle that the shareholders’ equity is locked into the company and may not be returned to the shareholders except as permitted by law, for example, by way of a lawful distribution, over which the collective corporate bodies have control, or a lawful reduction of capital (ditto). It may be that this problem arises only in the comparatively limited case where both P and Co have independent causes of action against D and D’s actions cause losses to each of them which to some extent overlap. But it still needs to be solved.
The above argument also explains why the exclusion rule does not apply when P sues as creditor (even if P is also a shareholder in Co). A shareholder is a member of the company and is taken to have entered the organisation on the basis that its assets should be under collective control and that shareholders’ contributions are returnable only in the permitted ways. This membership status justifies the excluded loss rule, even where there is nothing in the company’s articles explicitly constraining P’s powers to contract outside the company with people like D. The excluded loss rule arises out of the nature of the company as an organisation, rather in the way that Lord Wilberforce in Howard Smith Ltd v Ampol Petroleum Ltd ( AC 821, PC) deduced a restriction on directors’ powers from the structure of the articles as a whole rather than from the text of any particular article.
Making the distinction between shareholders and creditors does not require acceptance of Lord Reed’s proposition that ‘creditors will not suffer any loss so long as the company remains solvent’ , so that the extraction of assets from the company short of insolvency is of no concern to them. In the case of tradable debt, the value of the debt does fluctuate according, inter alia, to the amount of the asset cushion a company holds, even when it is not insolvent on either a balance sheet or cash flow test. The market is in the business of calculating the probabilities of repayment and if recovery in litigation by P acting as a creditor prevents recovery of significant assets by Co, the other creditors may well suffer a drop in the market value of their debt. The core point about creditors, as Lord Reed asserts, is that they do not become members of the company and so cannot be said to have entered into organisational obligations which restrict their capacity to recover overlapping losses.
Despite Lord Reed’s elegant elaboration and extension of the rationale underlying the first limb of the rule in Foss v Harbottle, there is something anachronistic about using that rule as the basis for addressing the problem of overlapping loss. After all, with one exception, the rule is no longer part of British law, having been overtaken by Part 11 of the Companies Act 2006. The one exception arises from the unaccountable failure of the Law Commission to think that ‘double derivative’ actions (where P wishes to enforce the rights of a subsidiary of the company of which P is a member) were worth bringing within the statutory regime. That mess was cleared up by Briggs J (as he then was) in Re Fort Gilkicker ( Ch 551) by resuscitating the Foss rule, but in that narrow context only. Was it a good idea to remove the stake from the heart of this corpse a second time in order to let it rise up and attack the overlapping loss problem?
The best reason for reforming the first limb of the rule in Foss v Harbottle was that, while it was very effective at preventing an individual shareholder from enforcing Co’s rights when there was available, theoretically, a superior body representing the company to take that decision, it did nothing (or very little) to ensure that that alternative decision-maker actually addressed its mind to the question. In broad terms, the individual could not enforce the company’s rights if the wrong in question was ratifiable by a simple majority in general meeting and, even if it was not, the individual could still not sue unless the alleged wrongdoers controlled the general meeting. Since many breaches of directors’ duties were ratifiable and all wrongs done to the company by outsiders in effect were (since the general meeting might release the company’s claims against them), the individual’s scope of action to enforce the company’s rights was very limited. It did not help P to gain locus standi that the general meeting of the company had not considered the matter and that there was no likelihood of its so doing.
In Foss v Harbottle itself Wigram V-C suggested the individual might have been allowed to proceed where ‘all means have been resorted to and found ineffectual to set [the general meeting] in motion’ (at 495). Had this point been taken up in later decisions, it might have been that an individual shareholder, at least in a company with highly dispersed shareholdings, having asked the board to convene a meeting of the shareholders to consider enforcement, might have been allowed to bring an action on behalf of the company if the directors’ reasons for refusing to call a shareholder meeting were not sustainable. But the point disappeared from sight. As it turned out, all the developed rule required of later courts was to examine whether the breach was (potentially) ratifiable (a question of law) and (possibly) the weight of the influence of the alleged wrongdoers in the hypothetical general meeting. Whether a general meeting would be summoned and what the non-involved shareholders would or might have thought about the issue, had it been put before them, and where the best interests of the company lay were not questions to trouble the court when denying the individual shareholder locus standi to enforce the company’s rights under the rule in Foss v Harbottle. Hence the criticism that the first limb of the Foss rule was more effective at denying the individual shareholder the right to sue than it was at ensuring the company’s rights were enforced in appropriate cases.
Under Part 11 of the Act, all this changed. Of course, one can debate how well the Act has turned out. However, it is clear that Part 11, following the example of other common law jurisdictions, was designed, perhaps too cautiously, to re-focus the courts’ attention on where the company’s best interests lay in relation to the prosecution of the litigation and on any evidence that might exist about what the non-involved shareholders of the company actually thought about the merits of initiating litigation. The court now reviewed the substance of the P’s application to enforce Co’s rights; P was no longer to be turned away on the basis that a general meeting might some day get around to considering the issue.
The relevance of this exegesis on the reform of the rule in Foss v Harbottle is that a similar problem arises in relation to the rule which excludes P from recovering for overlapping losses, even where the company has made no effort to recover its loss and it is not clear that it ever will. Like the first limb of the rule in Foss, the ban on the recovery by P of overlapping loss is more effective at preventing P from recovering than at ensuring that D does compensate the alternative claimant (Co) for the loss suffered. Good for D but not necessarily good for anyone else. In fact, under Lord Reed’s approach, that perverse consequence of the rule has been strengthened because he would do away with the exception referred to above and established by the Court of Appeal in Giles v Rhind ( Ch 618). This permitted P to sue for overlapping loss where Co’s incapacity to claim for it resulted from the very wrong committed by D, for example, by depriving Co of all its assets so that it could no longer give security for costs in litigation brought by it. In Lord Reed’s view the law simply does not recognise the overlapping loss as a loss suffered by P at all (at ) and so whether Co will or could sue to respect of it is irrelevant.
The minority view
The minority’s view that the overlapping loss rule should go altogether can be seen as a response to this unattractive aspect of even the modified rule excluding P from recovering this part of P’s loss. According to the minority, both P and Co should be able to sue and recover from D to the full extent of the losses inflicted on them by D’s unlawful actions. Any problems arising can be sorted out in the course of the litigation actually brought; there is no need to exclude P at the outset from recovering a particular head of loss, which, it might turn out, Co has taken and will take no steps to recover anyway.
The problems that might arise on this approach are two-fold and are inter-related: the risk of double recovery against D and the risk that Co will not be able to recover if P has got there first. One solution would be simply to remove D’s protection against double recovery in relation to overlapping loss claims. Co’s claim against D would not then have priority in the strong sense embodied in the Prudential case, but Co would be able to recover its loss in full from D even if P had already succeeded in full against D – assuming D was still worth suing. This is arguably all Co is entitled to. There are certainly some statements in the minority’s opinion which attach little weight to protecting D against the double recovery risk. For example, ‘...given the choice between ensuring that the claimant is fully compensated for the wrong done to him and eliminating any risk that the defendant might have to pay excessive compensation, I consider that the choice should be in favour of giving priority to protecting the interests of the innocent claimant rather than to giving priority to protecting the interests of the wrongdoing defendant’. (Lord Sales at ). But there is enough discussion in the minority’s opinion of the techniques for avoiding double recovery to suggest that they were not proposing a clean break with double recovery protection in cases of overlapping loss. One weakness of the minority opinion is that it does not develop a full analysis of how double recovery rules operate in this context. Do the minority envisage a world in which D is typically exposed to double recovery or one in which D will normally be protected from it, with exposure being the atypical case? If the latter is the position, how will the rules protecting D ensure that Co’s claim is given priority or do they think Lord Reed’s organisational arguments lack weight (as is suggested at ), so that priority for Co’s claim in unnecessary?
A second approach is to focus on identifying the loss suffered by P so that it is separate from the loss suffered by Co (ie, P’s claim contains no overlapping element). On this argument, there is no double recovery (because the losses do not overlap) and, for the same reason, no impairment of Co’s claim against D arising out of P’s claim for a separate loss against D (assuming D can meet both judgments). Once more, it is unclear whether the minority are advancing the argument that P’s claims escape the Prudential rule only when they can be presented as containing no overlapping element. Certainly, the minority stress that P’s loss is not necessarily identical to Co’s. Sometimes, it will be more; sometimes less, because the losses are measured in different ways. Co’s loss (in a situation like that which arose in Prudential) is the value of the assets removed plus certain consequential losses arising out of the deprivation (the extent of whose recovery will depend on the nature of the cause of action held by Co). P’s loss, at least where the shares are publicly traded, consists of the market’s re-assessment of the value of the expected future stream of distributions arising out the change in the composition of Co’s assets (replacement of physical assets by a claim against D for their recovery).
To repeat, the fact that Co’s and P’s losses are different does not falsify the proposition that P’s losses may reflect the loss suffered by Co or that recovery by Co may restore some or all of P’s losses. The second solution would amount to confining the recovery by P to the amount whereby P’s loss will not be made good by recovery by Co. An example might be where the market assessment is that during the period of deprivation Co lost a unique opportunity to make an investment which would have yielded super-normal profits, but the legal rules on remoteness of damage attached to the cause of action asserted by Co do not recognise this loss (or recognise it in full). At one point in his judgement Lord Sales (around ) appears to move towards the proposition that what P should recover from D is the amount of P’s loss above and beyond that suffered by the company. However, it is very doubtful that this is what the minority were suggesting. It is a line of argument which can probably be accommodated within the ‘reflective loss’ rationale of Johnson and so it does not explain the minority’s wholesale attack on the reasoning in that case. And it would require the court to engage in a highly sophisticated, even speculative, valuation exercise. We know from US securities class action litigation how difficult it is to isolate the impact upon the share price brought about by revelation of new facts about the company. To throw into that assessment an additional set of hypothetical facts about corporate recovery of the assets and the uses to which the recovered assets might have been put seems extremely unlikely to generate defensible decisions.
As Lord Sales remarked at the outset of his judgement, ‘The facts in this case are relatively simple. The legal issues are more complex’. (). In truth, those facts presented the Supreme Court with a very modern dilemma. On the one hand, it could prefer with the minority the approach to law-making often adopted by modern legislators. Do what is ‘fair’ or ‘just’ or ‘reasonable’ in the circumstances of the particular case and avoid setting down rules in advance which will apply across the board. All can be sorted out ex post. Reliance on such standards suits law-making in complex areas of social life where the best outcome (or limited range of outcomes) cannot be identified in advance. On the other, the majority’s view reflected a classical approach to company law, appropriately derived ultimately from a mid-nineteenth century decision, seeking to work out answers to difficult problems from an analysis of the legal structure of the company. Such an approach enhances the conceptual coherence of company law and makes life easier for courts. Since the only two judgements in favour of the majority view were given by Scottish judges, and despite the silent agreement of some the English (and Welsh) judges in the majority, perhaps it is not fanciful to detect an underlying contrast between English common lawyers’ commitment to getting the right answer on the facts of each case and Scottish legal training more open to civilian influences and a top-down approach.
Paul Davies QC (hon), FBA is Senior Research Fellow at the Centre for Commercial Law, Harris Manchester College, University of Oxford.
My thanks to Luca Enriques and Kristin van Zwieten for comments on an earlier version.