Halton International Inc & Anor v Guernroy Ltd

British Mediterranean Airways Ltd. (“BMed”) was founded in 1994. It operates scheduled services between London and various destinations in Asia and the Middle East. The parties are all shareholders. The first claimant (“Halton”) is a company controlled by Dr Tabbara. He and the second claimant, Mr Kaddoura, were two of the original founders of BMed. They are both Lebanese by birth, but with extensive world-wide business interests. The initial proposal for the airline came from another Lebanese businessman, Mr Lababidy, a friend and colleague of Dr Tabbara. The Defendant (“Guernroy”) is also a shareholder in BMed, its shares being held as nominee for various trusts connected with Wafic Said, a wealthy Syrian businessman. He was also approached initially to invest by Mr Lababidy.

BMed’s initial performance was poor and by 1996 it needed substantial fresh capital. At that time it was offered a possible solution to its problems, in the form of a franchise in the Middle East from British Airways (BA). By August 1996 a franchise agreement had been entered into with BA but it was dependent on BMed being able to raise £5.5 million of new equity. This proved an insuperable obstacle, until in October 1996 an arrangement was reached between the shareholders to enable Mr Said to seek to raise the necessary finance, including the so-called “voting agreement”.

The judge found that the voting agreement was signed by all the principal shareholders by 30^th October 1996. It had been sent to them on 23 October 1996 under cover of a memorandum from Mr Said’s lawyer explaining why it was necessary. This memorandum contained the following statement:

"Guernroy Limited will ensure that all of the existing shareholders of British Mediterranean are given the opportunity to participate on a pari passu basis in all financing proposed during the period of the authority described below.

The directors of Guernroy are of the opinion that in order to protect the interests of British Mediterranean, a full authority must be granted to Guernroy...”

The arrangements were to “remain in effect for so long as the British Airways franchise continues”.

In the voting agreement itself, the Preamble (para E) stated that the parties had agreed that “… voting control should be vested in Guernroy to facilitate negotiations with British Airways…” By Clause 1, headed “Transfer of Voting Rights”, the other shareholders appointed Guernroy “to be their true and lawful attorney” to take action and execute documents on their behalf “as Guernroy shall in its absolute discretion think proper to do…”, including voting on their behalf at any meetings. By clause 2.4, they were required to deliver to Guernroy all share certificates or other title documents. Clause 4 expressed the mutual intention that Guernroy should seek to procure funding to meet the requirements of British Airways, and should be free to do so -

4.2

…. in any manner which Guernroy, in its absolute discretion, considers fit including for the avoidance of doubt funding raised pursuant to or in connection with a fresh issue of Shares (whether by way of rights or otherwise) or other securities in the Company or by way of bank borrowings or other borrowings of any nature whatever….”

As to what followed the signing of the agreement, it is sufficient for present purposes to quote the summary from Mr Steinfeld’s skeleton:

“Thereafter, on 6 February 1997 at a short noticed EGM of BMed, Guernroy, who waived notice for the meeting on behalf of the Cs, used its powers under the voting agreement to ensure that 27.5 million new shares in BMed were issued at a price of 20p each to (a) Guernroy, which received 16.25 million of them; to (b) Lord Hesketh, the chairman of BMed’s board and a personal friend of Mr Said, who was allotted 5 million of them; and to (c) two other companies controlled by personal friends of Mr Said, who between them received the remaining 6.25 million shares. In particular, D used the voting agreement to pass a resolution disapplying the Cs’ rights of pre-emption under BMed’s articles. The effect of these allotments was to increase D’s holding in BMed from 20.1% to 50.9% and to reduce the holding of C1’s predecessors from 23.6% to 4.9%, and C2’s holding from 17.5% to 3.7%. The Judge has found as a fact that the Cs had on 29 November 1996 been sent a letter by D which gave them notice that investment in BMed shares was being invited by D but that they chose not to take part.

The Cs did not become aware of the allotments of February 1997 until September 2001. At that time there was a further issue of shares in BMed. The Cs were invited to participate on the basis of their holdings as diluted by the allotments of February 1997, which was how they discovered that their interest had been diluted. They decided not to participate at that stage because they did not wish to prejudice their case that the dilution of their interest was wrongful and that they should have been offered the opportunity to purchase shares based on their holding pre-6 February 1997.”

The present proceedings were issued on 29 October 2003. The claimants asserted that Guernroy had used the voting agreement to obtain a profit for itself, in that it obtained shares in 1997 and in 2001 that have greatly increased in value. They claimed that Guernroy owed a fiduciary duty to them not to use the voting agreement to obtain a profit for itself without their informed consent. The judge rejected the claim that Guernroy owed such duty, which, he said, “simply fails to recognise the commercial realities involved”:

“… Mr Said asked to be given carte blanche and that is what in terms Guernroy received. To superimpose on this a radically different set of obligations would be quite inconsistent with the relevant circumstances in which the voting agreement came to be made, as reflected in the terms of the agreement itself.”

He went on to hold that in any event the claim was time-barred, the proceedings having been commenced more than six years after the critical actions on 6^th February 1997.

The claimants sought permission to appeal on seven grounds, the last being the limitation issue. Chadwick LJ indicated that there were realistic prospects of success on (at least) some of the substantive grounds, but not on the limitation issue (ground (7)), which was fatal to the claim as a whole. He therefore refused permission, but indicated that if the application were to be renewed argument should be limited to the limitation issue, and that if permission were granted on that issue the court would proceed to determine it. Although attempts were made by the claimants to vary that order, they were rejected. It follows that the only issue before us is limitation.

There is a certain artificiality in seeking to decide whether a claim is time-barred, where the judge has been held that no valid claim exists, and where accordingly its nature (if it did exist) has not had to be analysed in detail. In the circumstances, as I see it, we must consider the issue in the most favourable light from the appellants’ point of view. Only if we are able to conclude, on that basis, that the appeal must fail, would we be justified in shutting out what has been held to be an arguable appeal on the other grounds. As we indicated to Mr Steinfeld at the outset, one possible outcome of this hearing might be that we simply give permission to appeal on the limitation ground, to be argued in due course in conjunction with the other grounds. However, since we have heard full and expert argument on this issue, and it is one of some general importance, we should attempt if possible to reach a concluded view.

Mr Steinfeld has helpfully summarised his case as follows in his most recent skeleton:

“…it is appropriate to consider what, in their barest essentials, the facts relevant to limitation are. On the Cs’ case these are that D held a power of attorney over Cs’ shares in BMed (“the voting agreement”). The immediate purpose of this was to enable D to exercise the votes that were attached to those shares (“the votes”).

For the purposes of this part of the appeal it must be assumed that (contrary to the findings of the Judge) D was under a fiduciary obligation to exercise those votes for the benefit of the Cs. By reason of that obligation it was prevented from exercising the votes for its own benefit without first obtaining the informed consent of the Cs. In breach of that fiduciary obligation it exercised the votes to allot new shares in BMed to itself (“the new shares”).

The Cs seek an order requiring D to make available to the Cs a portion of the new shares, based pro rata on the proportion of BMed that they owned before allotments, in return for payment of the cost price. The basis of this proprietary claim is that the shares when acquired by D were as to that portion held by it on constructive trust for the Cs.”

It is common ground that the claim is statute-barred, unless it can be brought within the exception provided by section 21(1) of the Limitation Act 1980:

“21(1) No period of limitation prescribed by this Act shall apply to an action by a beneficiary under a trust, being an action—

(a)

in respect of any fraud or fraudulent breach of trust to which the trustee was a party or privy; or

(b)

to recover from the trustee trust property or the proceeds of trust property in the possession of the trustee, or previously received by the trustee and converted to his use…”

By section 38, applying the Trustee Act 1925 s 68(17), trusts for these purposes include “implied and constructive trusts”. For these purposes, as was explained by Millett LJ in Paragon Finance plc v DB Thakerar & Co [1999] 1 AER 400, 408, it is necessary to distinguish between two possible uses of the term “constructive trust”:

“… the expressions 'constructive trust' and 'constructive trustee' have been used by equity lawyers to describe two entirely different situations. The first covers those cases… where the defendant, though not expressly appointed as trustee, has assumed the duties of a trustee by a lawful transaction which was independent of and preceded the breach of trust and is not impeached by the plaintiff. The second covers those cases where the trust obligation arises as a direct consequence of the unlawful transaction which is impeached by the plaintiff….”

Only the first category (“class 1”) is treated as a “trust” for the purposes of the section 21 exception to the ordinary limitation rules.

He explained the thinking behind the special treatment of class 1, by reference to the history of limitation rules as applied to trustees:

“Before 1890, when the Trustee Act 1888 came into operation, a claim against an express trustee was never barred by lapse of time. The Court of Chancery had developed the rule that, in the absence of laches or acquiescence, such a trustee was accountable without limit of time. The rule was confirmed by Section 25(3) of the Judicature Act 1873, which provided that no claim by a cestui que trust against his trustee for any property held on an express trust, or in respect of any breach of such trust, should be held to be barred by any statute of limitation.

The explanation for the rule was that the possession of an express trustee is never in virtue of any right of his own but is taken from the first for and on behalf of the beneficiaries. His possession was consequently treated as the possession of the beneficiaries, with the result that time did not run in his favour against them: see the classic judgment of Lord Redesdale in Hovenden v. Lord Annesley (1806) 2 Sch. & Lef. 607 at pp. 633-4.

The rule did not depend upon the nature of the trustee's appointment, and it was applied to trustees de son tort and to directors and other fiduciaries who, though not strictly trustees, were in an analogous position and who abused the trust and confidence reposed in them to obtain their principal's property for themselves. Such persons are properly described as constructive trustees.” (p 408 f-j)

He rejected the argument that this distinction had not survived the Limitation Act 1939.

The distinction drawn by Millett LJ had its precursor in a judgment of the Privy Council in Taylor v Davies [1920] AC 636 (considering an Ontario limitation statute). In that case, the defendant was an “inspector”, appointed under an insolvency statute, to oversee assignments for the benefit of creditors. Although, by virtue of his position, he could not rely upon a transfer made to him without his having made full disclosure to the creditors, it was held that he had a limitation defence because he was not a "trustee" of the property within the meaning of the relevant statute.Viscount Cave explained the background of the exception, and its extension to those who, though not named as trustees, had “assumed the position of a trustee for others or had taken possession or control of the property on their behalf”:

"… These persons, though not originally trustees, had taken upon themselves the custody and administration of property on behalf of others; and though sometimes referred to as constructive trustees, they were, in fact, actual trustees, though not so named.It followed that their possession also was treated as the possession of the persons for whom they acted,and they, like express trustees, were disabled from taking advantage of the time bar…." (p 651)

The effect of that decision was summarised by the Privy Council in Clarkson v Davies [1923] AC 100, 110:

“… it was there laid down that there is a distinction between a trust which arises before the occurrence of the transaction impeached and cases which arise only by reason of that transaction.”

To set the scene for the discussion in the present case, it is helpful to compare (as the judge did) two recent cases in this court, which were held to fall either side of the line: JJ Harrison (Properties) Ltd v. Harrison [2002] 1 BCLC 162, and Gwembe Valley Development Co. Ltd. v. Koshy [2004] 1 BCLC 131.

In Harrison the claimant was a family property company, of which the defendant was a director. In 1985 a property owned by the company was valued at £8,400, but in a side letter the valuer said that it “may have some development potential” and that the valuation did not take this into account. The side letter was not disclosed to the company, but the director was aware of it. In 1986 he bought the property from the company without disclosing the side letter, or the fact that he had himself made an application for planning permission. He later sold it for a very substantial profit. More than six years later the company sued him for an account of the proceeds of sale of the land. His limitation defence failed in the Court of Appeal on the ground that the claim was within class 1.

In Gwembe Valley, the defendant was managing director of the claimant company, GVDC, and a shareholder in it. He also owned a majority of the shares in and controlled another company, Lasco, which lent money to GVDC. He arranged the loans. As a result of the loans GVDC acknowledged a debt to Lasco of U$5.8m. This was on the basis that the sum advanced, 56.4m kwacha (Zambian currency) was equivalent to U$5.8m at the official exchange rate. However, the kwacha advanced to GVDC by Lasco had been obtained for only about U$1m by the defendant. The case against the defendant was that he did not disclose to the board of GVDC his interest in Lasco, or the profit that Lasco made on the loans to GVDC. It was held that, apart from fraud, the claim would have been time-barred, because his liability to account for the secret profit was not within class 1, and therefore the exception provided by section 21(1)(b) did not apply. (The claim ultimately succeeded, on the basis that the defendant had fraudulently concealed from GVDC his interest in Lasco and the true cost of the kwacha lent to GVDC.)

The difference between the two cases, in short, was that while in the former the director had a pre-existing “trustee-like responsibility” in relation to the particular property which was the subject of the action, in the latter he did not. As Mummery LJ explained:

“… The judge made no finding that any payments made by GVDC to Lasco, or any payments made out of the current account, were improperly made. The only thing wrong with them was Mr Koshy’s failure, in his separate capacity as director of GVDC, to make full disclosure to that Board of the nature and extent of his own financial interest

…

If that is the correct analysis, then it is clear in our view that any trust imposed on Mr Koshy is a class 2 trust, within Millett J’s classification … [In Harrison] the director transferred to himself property which had previously belonged to the company, and in relation to which he had ‘trustee-like responsibilities’ before the transaction in question. By contrast, Mr Koshy’s liability to account for undisclosed profits, and any constructive trust imposed on those profits, do not depend on any pre-existing responsibility for any property of the company. They arose directly out of the transaction which gave rise to those profits, and the circumstances in which it was made. The fact that Mr Koshy was in a pre-existing fiduciary relationship with the company was not enough, by itself, to bring the case within class 1…” (p 165 paras [118]-[119])

In the present case, Patten J followed Gwembe in holding that the present claim fell within class 2. He said:

“….On the Claimants' case (Guernroy) acquired the shares through its own breach of duty in circumstances which give rise to what amounts to a remedial constructive trust. The case can be distinguished from the position of (eg) a director of a company who controls the property of the company and owes pre-existing duties to the company in respect of it:seeJJ Harrison (Properties) Ltd vHarrison[2002] 1 BCLC 162.Guernroy owed no duties to anyone in respect of the unissued share capital of BMed. The company issued the shares in return for the issue price which was paid. The Claimants' case is that the acquisition of the shares constituted a breach of duty to the existing shareholders, but it is not alleged that Guernroy in any sense held the unissued shares for the Claimants prior to the alleged breach. The most that can be said is that it owed fiduciary duties to the Claimants in respect of the voting powers and that it is through the alleged misuse of those powers that the shares have been acquired. In my judgment, this brings the case within class 2: see Gwembe Valley Development Co. Ltd v Koshy [2004] 1 BCLC 131 at p.165 g-i.”

Before us, conscious no doubt of the need to identify something more than a mere pre-existing fiduciary relationship, Mr Steinfeld has placed the emphasis of his argument on the “specific property” represented by the voting rights, which were in effect transferred to Guernroy. He puts it thus in his skeleton:

“(In Gwembe), although there was a fiduciary relationship between the parties prior to the transaction, it did not attach to any particular property. It was merely a relationship that gave rise to a liability to account once the transaction giving rise to the profit had occurred. In the present case D had powers in relation to specific property of the Cs, namely the voting rights attached to the Cs’ shares in BMed, that pre-existed the transaction in question and which were themselves the means to obtain the further property (the shares obtained in 1997 and 2001) that is held on trust for the Cs. The voting rights constituted in that sense property held under a pre-existing trust relationship which were as much the means of enabling D to obtain the new shares as the money subscribed for those shares. Had the money itself been held on trust for Cs the shares so acquired would undoubtedly be trust property within category 1. There is no reason why the result should be any different where the means of acquisition was the use of the voting rights when those rights were themselves held on trust for the Cs.” (emphasis added)

As an alternative, he criticises the judge’s approach for giving undue weight to the particular characteristics of the property acquired, rather than the source of the obligation:

“… the effect of the Judge’s approach is to regard the character of the property obtained by D as determinative of the class of trust to which it is subject in D’s hands. It is submitted that this is a red herring. The fact that the new shares did not come into existence save through the transaction impeached by the Cs is a mere accidental characteristic of the property that D obtained. If, for example, a director, in breach of his fiduciary duty, used his powers as director to obtain for himself shares in a subsidiary he would hold them on a class 1 trust. He would not be able to benefit from the Limitation Act by saying that he owed no duty in respect of the shares before he procured them to be issued to himself. Why should D be in a different position from such a director?”

Neither way of putting the point is readily apparent from the judgment. The first implies a particular interpretation of the agreement which is itself controversial. Mr Girolami, for the respondent, says that the agreement on its face conferred no more than a power of attorney, and that it is a distortion to speak of the voting rights being held “on trust for the claimants”. There is a further potential difficulty of causation. Arguably the use of the claimants’ votes to disapply their pre-emption rights had no practical effect, since, as the judge appears to have found (paras 156-159), the claimants would not have wished to exercise them.

However, adopting the approach outlined earlier, I think it right to proceed at this stage on the footing that Mr Steinfeld’s present formulations are consistent at least with a possible interpretation of the facts, and that they represent the strongest ways of putting the case from the appellants’ point of view.

Section 21(1) provides an exception to the ordinary limitation rule that civil actions are barred after six years. Such an exception needs to be clearly justified by reference to the statutory language and the policy behind it. It is important therefore to keep in mind the reasoning behind the exception. It is not about culpability as such; fraud may not be sufficient to avoid the ordinary rule. (Footnote: ¹) It is about deemed possession: the fiction that the possession of a property by a trustee is treated from the outset as that of the beneficiary. In the words of Millett LJ, the possession of the trustee is “taken from the first for and on behalf of the beneficiaries” and is “consequently treated as the possession of the beneficiaries”. An action by the beneficiary to recover that property is not time-barred, because in legal theory it has been in his possession throughout.

Turning to the present case, to treat the voting rights as class 1 trust property, even if justified on the facts, does not assist the claim. That is not the property which the claimants are seeking to recover. The most that can be said is that the voting rights were part of the means by which the defendants acquired the new shares. Mr Steinfeld was unable to point to any case in which that has been sufficient for the purposes of section 21 or its predecessors. For the exception to apply there must be a trust (or trustee-like responsibility) for specific existing property, not merely for the means to obtain it in the future.

To overcome this problem Mr Steinfeld sought to draw an analogy with the old case of Keech v Sandford (1726) Sel.Cas Ch. 61. In that case the trustee of a lease applied for a renewal of the lease for an infant beneficiary, and, on this being refused, obtained a renewal for himself; it was held that he held the lease on trust for the infant. Mr Steinfeld suggests that in such a case the trust of the renewed lease would be a Class 1 trust, even though it arose from the impugned transaction, and the earlier lease was no more than the means to obtain it. He noted that in the leading case of Regal (Hastings) v Gulliver [1967] 2 AC 134, 138, Keech v Sandford was treated as an example of the general rule of equity that a person in a fiduciary position cannot enter into engagements in which his interests may conflict with the interests of those he is bound to protect (p 137-8).

Neither of those cases was concerned with the limitation rules. The rule (exemplified by Regal), which prohibits a fiduciary from making a secret profit is not in itself an accurate guide to the scope of class 1. That much is evident from Paragon. Keech v Sandford might arguably be brought within class 1, but, if so, only because of the special nature of the property involved. As was explained in Biss v Biss [1903] 2 Ch 40, 56 per Collins MR, the renewal is treated as “an accretion to or graft upon the original term arising out of the goodwill or quasi-tenant right annexed thereto”. It provides no analogy for a similar link between the voting rights and the new shares in the present case.

As to the alternative argument, the way in which the new shares came into existence cannot be dismissed as “a mere accidental characteristic”. That was an essential feature of the transaction. The dividing line drawn by the cases is between a trust which arises “before the occurrence of the transaction impeached” and one which arises “only by reason of that transaction”. In this case, as the argument in effect concedes, the new shares came into existence “only by reason of the transaction impeached”.

For these reasons, I would dismiss the appeal.

I agree.

I also agree.