United Kingdom Supreme Court

Michaelmas Term

[2017] UKSC 68

On appeal from: [2016] EWCA Civ 708

LORD NEUBERGER: (with whom Lord Clarke, Lord Sumption and Lord Hodge agree)

1.                   This appeal raises the issue whether the daily vessel-operating expenses of shipowners incurred while they were negotiating to reduce the ransom demands of pirates should be allowed in general average - ie whether those expenses should be shared proportionately between all those whose property and entitlements were imperilled as a result of that seizure - or whether they must be borne by the shipowner alone.

2.                   General average refers to the system of maritime law by which sacrifices of property made, and loss and expenditure incurred, as a direct result of actions taken for the purpose of preserving a common maritime adventure from peril are rateably shared between all those whose property is at risk in the adventure. The principle of rateable sharing of such losses between parties to a maritime adventure appears to date back at least to the law of the Rhodians. Having been adopted by the Romans, it passed on a customary basis into European sea laws of the Middle Ages, and thence into modern European Codes. It appears that the expression “general average” started to be used in English judgments around the end of the 18th century and was first authoritatively discussed judicially in this country by Lawrence J in Birkley v Presgrave (1801) 1 East 220, 228-229. It was first recognised statutorily in section 66 of the Marine Insurance Act 1906.

3.                   The York-Antwerp Rules are an internationally agreed sets of rules, the first set (under that name) propounded in 1877, since when they have gone through a number of versions. The latest version was agreed in 2016. The Rules are designed to achieve uniformity in ascertaining which losses fall within the principle, in determining the method of calculating those losses, and in deciding how they are to be shared. Although internationally agreed between relevant expert and interested bodies, the Rules are not the subject of English legislation or international convention, and they derive legal force only through contractual incorporation. In the present case the 1974 version of the Rules was contractually incorporated into the relevant carriage contract. I will refer to that version as “the Rules”. The Rules are in English and French, and for the most part I shall confine myself to the English version.

4.                   The Rules are introduced by a “Rule of Interpretation”, which states that:

5.                   Three of the lettered Rules are of particular relevance to this appeal, namely Rules A, C, and F. Rule A is in these terms:

6.                  The numbered Rules play no part in these proceedings, save that some reliance has been placed on Rule XI. That Rule is concerned with “Wages and Maintenance of Crew and other expenses bearing up for and in a Port of Refuge etc”, and is the second longest of the Rules. It provides among other things for crew wages and maintenance to be recoverable in general average where Rule XI applies.

The factual background

7.                   On 29 January 2009 the chemical carrier MV Longchamp (“the vessel”) was transiting the Gulf of Aden on a voyage from Rafnes, Norway, to Go Dau, Vietnam, laden with a cargo of 2,728.732 metric tons of Vinyl Chloride Monomer in bulk (“the cargo”). The cargo was carried under a bill of lading dated 6 January 2009 which stated on its face that “General Average, if any, shall be settled in accordance with the York-Antwerp Rules 1974”.

8.                   At 06.40, seven heavily armed pirates boarded the vessel. The pirates commanded the master to alter course towards the bay of Eyl, Somalia, where she arrived and dropped anchor at 10.36 on 31 January 2009. At 14.05 on 30 January 2009 a negotiator for the pirates boarded the vessel and demanded a ransom of US$6m. The vessel’s owners (“the owners”) had meanwhile formed a crisis management team who had set a target settlement figure of US$1.5m. On 2 February 2009 an initial offer of US$373,000 was put to the pirates. Negotiations between the pirates’ negotiators and the owners’ crisis management team continued over the following seven weeks with various offers and counter-offers being made.

9.                   Eventually on 22 March 2009, after a negotiation period of 51 days, a ransom was agreed in the amount of US$1.85m. On 27 March 2009 the ransom sum was delivered by being dropped at sea. At 07.36 on 28 March 2009 the pirates disembarked and at 08.00 that day the vessel continued her voyage.

10.              It is accepted that the US$1.85m ransom payment itself can be allowed under Rule A. It is also accepted that the costs and expenses of the negotiator in relation to the ransom, Captain Ganz, and the costs and expenses of his special advisers, NYA International, are allowable. There was a dispute about the allowability of a sum of around US$20,640 in respect of media expenses but that is no longer challenged by the cargo interests.

11.              The essential issue on this appeal is whether the vessel-operating expenses incurred during the period of negotiation (“the negotiation period expenses”) are allowable in general average under Rule F. Those sums are:

12.              The average adjuster, Mr Robin Aggersbury of Stichling Hahn Hilbrich, considered that the negotiation period expenses were allowable under Rule F on the basis that they were incurred “during a negotiation period of about 51 days” which enabled “an amount of US$4,150,000 [to be] saved in the common interest of all property owners concerned, which would otherwise have been recoverable as per Rule A”. The 51-day period to which he referred was, as explained above, from 30 January to 22 March 2009. The cargo was valued at destination at US$787,186 and the value of the vessel was assessed at US$3,947,096; so cargo interests were liable for 14.44% of the total general average expenditure. Following publication of the adjustment, the cargo interests requested and obtained a report (“the Report”) from the Advisory Committee of the Association of Average Adjusters. The Report set out the facts in considerable detail, and concluded, by a majority of four members to one, that the negotiation period expenses did not fall within Rule F.

13.              The cargo interests had previously made payments on an account of general average, but following the publication of the adjustment they issued proceedings challenging (in accordance with the Report) the adjuster’s conclusion that the negotiation period expenses fell within Rule F, and seeking an appropriate repayment.

14.              The owners’ argument involves the following steps. First, it is rightly common ground that the US$1.85m ransom paid to the pirates for the release of the vessel was expenditure which was “a general average act” within Rule A. Secondly, the negotiation period expenses claimed fell within the expression “expense incurred” by the owners within Rule F. Thirdly, those expenses were “incurred in place of another expense”, namely the US$4.15m saved as a result of the negotiations. Fourthly, those expenses, being US$160,000, are less than “the general average expense avoided”, namely the US$4.15m (and for the sake of simplicity I will treat this as the saving, although the actual saving was somewhat less by virtue of expenses such as those paid to Captain Ganz and NYA). Fifthly, it follows from this that the negotiation period expenses are properly allowable under Rule F.

16.              The cargo interests’ first contention is based on the proposition that it would not have been reasonable for the owners to have accepted the pirates’ initial ransom demand for US$6m. On that basis, it is said that a payment of US$6m (or, more accurately, the saving of US$4.15m) would not have been “expenditure … reasonably … incurred” within Rule A, and therefore cannot qualify as an “expense which would have been allowable as general average” in Rule F. The judge accepted that, in order to succeed in its claim under Rule F, the owners would have to establish that it would have been reasonable for them to have accepted the pirates’ initial demand, but decided that, in all the circumstances, it would have been reasonable for the owners to have paid US$6m ransom. The Court of Appeal agreed with the judge’s analysis of the legal position, and declined to interfere with his conclusion that it would have been reasonable of the owners to have met the pirates’ initial demand.

17.              It is a difficult question whether the Court of Appeal ought to have concluded that the judge was entitled to conclude that it would have been reasonable for the ship-owner to have paid the pirates the US$6m which they initially demanded. While an appellate court should be slow to interfere with a trial judge’s finding of fact, this was not a finding of primary fact. And, at least on the face of it, one would have thought that it would have required very unusual circumstances for a ship-owner not to try and negotiate with pirates who had made such a very high demand. Further, the evidence suggests that no ship-owner accepted an initial demand made by Somali pirates and that their demands were generally pitched on the basis that they would be substantially reduced by negotiation. On the other hand, one must beware of the perils of wisdom of hindsight, and it is right to bear in mind that there was a wounded sailor on the vessel and that the cargo was perishable.

18.              In my opinion, it is not necessary to resolve this difficult issue, because I do not consider that the judge or the Court of Appeal were correct in assuming that the owners had to establish that it would have been reasonable to accept the pirates’ initial demand in order to justify the contention that the negotiation period expenses were allowable under Rule F. One does not need to examine the wording of the Rules to appreciate that the assumption made by the courts below would lead to very odd results, as explained by Hamblen LJ at [2016] Bus LR 1285, paras 62 to 64. It would mean that, if a ship-owner incurs an expense to avoid paying a reasonable sum, he can in principle recover under Rule F, whereas if he incurs expense to avoid paying an unreasonable sum (ie a larger sum), he cannot recover. The more obvious his duty to mitigate, and the greater the likely benefits of such mitigation, the less likely he would be to be able to recover. Such a state of affairs (apparently known to cognoscenti as the “Hudson conundrum”, after the writer who first described it) would be a remarkable result. Fortunately, examination of the wording of Rules A, C and F shows that it does not arise.

19.              Where I part company with the judge and the Court of Appeal is in relation to their view that the reference in Rule F to “another expense which would have been allowable as general average” is to an expense whose quantum is such that it would have qualified as a claim under Rule A. In my opinion, the reference to an “expense which would have been allowable” is to an expense of a nature which would have been allowable. First, the word “allowable” in Rule F naturally takes one to Rule C, where the similar word “allowed” is used, rather than Rule A, where there is no reference to anything being “allowed” (the same point applies to the French version - “admissible” in Rule F and “admis” in Rule C). Unlike Rule A, Rule C is concerned purely with the type of expense, and not with quantum. Secondly, the opening part of Rule F is unlikely to be concerned with quantum, as that is dealt with in the closing part, which imposes a cap on a sum recoverable under Rule F, namely “only up to the amount of the general average expense avoided”. Thirdly, the interpretation assumed in the courts below imposes an unnecessary fetter on the allowability of an “extra expense”, as there is already a reasonable fetter in the concluding part of Rule F. Fourthly, the interpretation I favour produces an entirely rational outcome: whenever an expense is incurred to avoid a sum of a type which would be allowable, that expense would be allowable, but only to the extent that it does not exceed the sum avoided.

20.             Applying that reasoning to this case, and subject to the discussion below as to the cargo interests’ other arguments, the US$160,000 falls within Rule F. The US$160,000 was incurred in order to avoid paying a US$6m ransom (or, more accurately, a ransom of around US$4m more than the ransom actually paid), and as the ransom was an allowable expense in principle, the US$160,000 therefore falls within Rule F, subject to the appellant establishing that it would have been reasonable to have paid a ransom of around US$2.4m (ie the ransom it did pay plus the US$160,000 together with the further expenses such as those paid to Captain Ganz and NYA). If the judge was even arguably entitled to reach the conclusion that paying a US$6m ransom was reasonable, it must have been reasonable to pay a ransom well under half that figure.

21.              Even if the analysis in para 19 above were not right, I would have reached the same conclusion. As pointed out by Lord Sumption in the course of the argument, where an unreasonably high sum is expended, there would be no reason not to hold that Rule F applied, albeit only to the extent of a reasonable sum, on the basis that the greater includes the less. Thus, if (contrary to the analysis in para 19 above), Rule F only applied where a sum was reasonably incurred, and in this case the judge had concluded that the maximum reasonable ransom would have been US$4m, then Rule F would have applied to US$4m of the US$6m ransom.

22.              I turn then to the second contention raised by cargo interests, which was the ground on which they succeeded in the Court of Appeal. That contention is that the negotiation period expenses do not fall within Rule F, because the payment of a reduced ransom of US$1.85m was not an “alternative course of action” to the payment of the ransom originally demanded, namely US$6m: it was merely a variant. This contention involves arguing that to trigger Rule F, it is not enough for a claimant to incur expense in achieving a result which costs less than what an allowable item would otherwise have cost: the expense must be incurred to achieve a result which involves replacing that allowable item with a different and cheaper item. As Lord Mance expressed it during the argument, this argument involves saying that Rule F applies only where some means is adopted to complete the adventure, and that means is different from that which might normally be expected.

23.              The notion that Rule F is only engaged in a case where the claimant achieves an “alternative course of action” in that sense was said by Hamblen LJ at [2016] Bus LR 1285, paras 38 to 40 to be supported by passages in the two leading books in English on general average. In paras F.01 and F.29, the editors of Lowndes & Rudolf, The Law of General Average and the York-Antwerp Rules, 14th ed (2013) write:

“As the name implies, substituted expenses are the expenses incurred in respect of a course of action undertaken as an alternative to - or in substitution for - the expense that would be allowable as general average.

For this rule to have any application there must have been an alternative course which, if adopted, would have involved expenditure which could properly be charged to general average.”

In Hudson & Harvey, The York-Antwerp Rules: The Principles and Practice of General Average Adjustment, 3rd ed (2010), para 11.33, there is this:

“Although Rule F is phrased in terms which refer to the incurring of the expense, its application in practice presupposes a choice between two (and sometimes more) different courses of action.”

24.              I am not convinced that, as a matter of language, those passages support the conclusion that Rule F can only be invoked when the claimant has taken an “alternative course of action”, but I accept that the prevailing view among the writers on the subject, and among those who work in the field, of general average may well be as Hamblen LJ suggested. Thus, it certainly seems to have been assumed to be the generally accepted position by Hoffmann LJ in his striking dissenting judgment in Marida Ltd v Oswal Steel (The Bijela) [1993] 1 Lloyd’s Rep 411, where, at p 423 he quotes with approval a passage from the 11th ed (1990) of Lowndes & Rudolf which is identical to that quoted by Hamblen LJ from the 14th ed.

25.              However, the law cannot be decided by what is understood among writers and practitioners in the relevant field (or even by views expressed by Hoffmann LJ in a dissenting judgment, especially in a case where the point did not strictly arise and does not appear to have been argued). Experience shows that in many areas of practical and professional endeavour generally accepted points of principle and practice, when tested in court, sometimes turn out to be unsustainable. I accept that it may be right for a court to have regard to practices which have developed and principles which have been adopted by practitioners, but they cannot determine the outcome when the issue is ultimately one of law. Further, as the opinions of the average adjuster and of the majority of the Advisory Committee of the Association of Average Adjusters in this case demonstrate, there is certainly no question of there being a universal view on the issue.

26.              Turning to the language of Rule F, I consider that this “alternative course of action” contention goes nowhere. Even if one accepts that the “extra expense” must involve an alternative course of action, it seems to me that the owners’ claim satisfies that requirement. It appears to me that (ignoring other sums for present purposes) the right analysis of the owners’ claim is that it is for (i) US$1.85m under Rule A and (ii) US$160,000 under Rule F, on the basis that (i) the US$1.85m, as a reasonable sum paid to ransom the vessel and the cargo, is admittedly within Rule A, and (ii) the US$160,000, as negotiation period expenses, represents “extra expense incurred in place of” the US$4.15m, the amount by which the ransom was reduced. On that basis, as I see it, the incurring of the US$160,000 did represent an alternative course of action, in the sense that the cargo interests use that expression, from the payment of the US$4.15m: the former involved incurring vessel-operating expenses whereas the latter involved paying a ransom.

27.              There is an alternative analysis of the owners’ claim, which is that it should be treated as being for a single sum of US$2.01m, namely the US$1.85m ransom actually paid plus the US$160,000 negotiation period expenses, under Rule F on the basis that this combined sum was “extra expense incurred in place of” the US$6m originally demanded. However, I do not see how that helps the cargo interests. Logically, their argument on this basis should be that the US$1.85m is disallowable under Rule F as it was not an alternative course of action from paying the originally demanded US$6m ransom, but the negotiation period expenses are recoverable under Rule F, as they did involve an “alternative course of action” - which is precisely the opposite of the cargo interests’ actual case, and indeed a nonsensical result.

28.              Accordingly, the cargo interests’ second contention cannot simply be based on the wording of Rule F. Their contention, as I see it, must be that the expenses incurred in negotiating a reduction in the cost of an allowable item do not fall within Rule F because the reduction in the cost of an allowable item which would be paid for anyway, and which falls within Rule A, cannot be within the scope of Rule F. I do not find it easy to see how one can get that out of the words of Rule A or Rule F. I suppose that one could take the analysis in para 27 above and argue that it works perfectly well where, as a result of the negotiation, an “alternative course of action”, within the restrictive use of that expression as urged by the cargo interests, was taken. However, given the problem identified in para 27 above with such an approach where the reduced sum is not such an “alternative course of action”, I am very dubious whether the approach can be justified anyway.

29.              Given that the Rules represent an international arrangement, it is particularly inappropriate to adopt an approach to their interpretation which involves reading in any words or qualification. As already mentioned, it appears to me that, as a matter of ordinary language, Rule F applies to the negotiation period expenses for the reasons given in para 26 above. To imply some qualification such as the requirement that those expenses must have been incurred so as to achieve an “alternative course of action” appears to me to be very dangerous. In the same way as an international convention or treaty, the Rules should be interpreted by a United Kingdom court “unconstrained by technical rules of English law, or by English legal precedent, but on broad principles of general acceptation”, to quote Lord Wilberforce in James Buchanan & Co Ltd v Babco Forwarding & Shipping (UK) Ltd [1978] AC 141, 152. As Lord Hobhouse said in King v Bristow Helicopters Ltd [2002] 2 AC 628, para 148, in relation to an article in the Warsaw Convention, “it is the unadorned language of the article to which attention must be directed”.

30.              Quite apart from this, the cargo interests’ second contention appears to me to lead to difficulties and potential anomalies in practice. Thus, there would be difficulties about deciding whether a particular variant was an “alternative course of action”. Towage to destination, extra costs of arranging dry-docking with cargo on board, overtime worked on repair or cargo operations (at least sometimes), and (historically) air freight instead of sea freight for spare parts were examples given by the respondents of “alternative courses of action” (mostly taken from Lowndes & Rudolf, op cit). But it is hard to see where the line is to be drawn. The difficulties about deciding whether overtime payments qualify is plain from reading Lowndes & Rudolf, op cit, para F.25; in addition, overtime payments are enhanced payments for the same work whose cost would have been recoverable under Rule A in any event. And if air freight can qualify if it is incurred instead of sea freight, it is hard to see much logic in disqualifying sea freight at a lower rate negotiated with a new party on a different type of vessel. Further, given that, on the cargo interests’ case, negotiation period expenses could not be claimed if they were incurred as a result of negotiating a reduction in the cost of repair with one shipyard, what would the position be if the negotiations were with a competing shipyard and/or in respect of a novel and different way of effecting the repairs?

31.              It also appears to me to be somewhat inconsistent in terms of logic that (as has been agreed between the parties in this case) the costs and expenses of Captain Ganz and NYA are subject to general average whereas the negotiation period expenses are not. It is clear that the costs and expenses claimed by and paid to Captain Ganz and NYA included costs and expenses attributable to the negotiations with the pirates (for instance, hotel bills for most of the 51-day period). They can only be justified on the basis that they were referable to the negotiations to reduce the ransom, in the sense that they were incurred solely because of the negotiations taking place. Accordingly, if they are claimable, it is hard to understand why the negotiation period expenses should not also be claimable.

32.              At [2016] Bus LR 1285, para 47, Hamblen LJ suggested that there were a number of anomalies if the negotiation period expenses were allowable. First, he mentioned the difficulty of establishing that the expenses would not have been incurred even if the initial demand for US$6m ransom had been accepted. I doubt that that problem would arise in most cases where the vessel-operating costs are said to fall within Rule F, and it may well arise in some cases where it would be common ground that Rule F would apply. In any event, it is for the claimant in each case to establish, on the balance of probabilities, that the delay caused by the negotiation would not have occurred if there had been no negotiation. Secondly, he said that, in a case such as this there could be no entitlement to claim vessel-operating costs as Rule F expenses until a demand had been made. I agree, but fail to see why it is an anomaly. Thirdly, Hamblen LJ pointed out that, in the absence of a demand, eg if there were simply negotiations, it might be hard to say when, or even whether, Rule F was engaged. I accept that there may be ransom cases where it is hard to determine at what precise point Rule F is engaged, but it would, I think, be a rare case where at some point early in the negotiations the pirates did not come up with a figure. Anyway, I suspect that point could apply to cases where Rule F is undoubtedly engaged. Quite apart from that, I do not accept that the fact that there may be difficulties for claimants in a few other ransom cases is a reason for holding that Rule F is not engaged in this case. More broadly, if (as appears to me to be appropriate) one views Rule F simply as entitling a claimant to claim in respect of an expense successfully incurred for the purpose of mitigating a loss, it seems to me that none of these points should give rise to concerns.

The cargo interests’ other arguments

33.              The cargo interests’ third contention is that, in order to be recoverable under Rule F, the negotiation period expenses must be shown to have been consciously and intentionally incurred by the owners, and there was no evidence that the owners or their agent had consciously decided to incur those expenses in order to reduce the ransom payable to the pirates. Indeed, Hamblen LJ said at [2016] Bus LR 1285, para 43 that “it does not appear that the owners ever considered that they faced a choice” and that there was “no evidence to suggest that they ever considered choosing between paying the ransom on demand and paying a lesser sum following negotiation”. Accordingly, runs the argument, the owners cannot recover under Rule F as they never made a conscious choice between paying the US$6m ransom initially demanded by the pirates or negotiating with the pirates.

34.              I do not accept that contention. The question whether one expense has been incurred “in place of another expense” must be assessed objectively. In this case, it is clear (and must have been clear at the time) that negotiations were (and would be) needed if the ransom was to be reduced, that such negotiations took (and would take) time, and that the passage of time resulted in the negotiation period expenses (and would result in expenses of that nature) being incurred. As the negotiations resulted in the ransom being reduced, it seems to me that, subject to any other argument, it must follow that the expenses incurred as a result of those negotiations were incurred “in place of” the US$4.15m saved (or that the expenses incurred plus the US$1.85m actual ransom were incurred “in place of” the original US$6m ransom demand).

35.              The cargo interests further contend that the negotiation period expenses were not “extra expense” within the meaning of that word in Rule F. This contention is based on the proposition that, in order to qualify as “extra expense”, an expense would have to be of a nature which would not normally have been incurred in response to the peril threatening the adventure. I can see no reason for giving the word “extra” such a restrictive meaning. First, it is not its natural contextual meaning, which, in my view, is simply an expense which would not otherwise have been incurred (but for the saving of the “other expense”). Secondly, such a meaning is supported by the contrast with the word “extraordinary” in Rule A. Thirdly, such a restrictive meaning lies unhappily with the French equivalent adjective, which is “supplémentaire”. I take some comfort from, but do not rely on, the fact that the word “extra” in Rule F has now been replaced, in later versions of the Rules, by the word “additional”.

36.              The cargo interests’ next contention is that the delay which led to the negotiation period expenses may well have occurred even if the owners had agreed to the pirates’ initial demand of US$6m. For instance, if the owners had accepted the US$6m, the pirates may have thought that they had pitched their initial demand too low, and would have increased it, leading to further negotiations and consequent delay. That is of course a possibility. However, it is inherent in the judge’s conclusion that he considered it more likely than not that the vessel and cargo would have been released promptly if the US$6m ransom demand had been accepted and paid. That was the sort of finding (albeit an implied finding, but necessarily so, in his conclusion) with which an appellate court should be very slow to interfere. And in this case it appears to me that we should clearly not question it: it was an eminently defensible finding. It is clear that a delay of some period would be inevitable as a result of the negotiations, and it is clear that the 51 days (between the initial demand of US$6m and the final agreement at US$1.85m) was inevitable as a result of the negotiations; on the other hand, to put it at its very lowest, it is not unlikely that none of the 51 days delay would have been suffered if the US$6m demand had been met.

37.              The cargo interests’ final contention is that, as Rule C excludes from general average expenditure which is an “indirect loss” including demurrage, and/or because Rule XI includes crew wages and maintenance where it applies, the claim in the present case must fail. In my opinion, there is nothing in that point. I accept that the negotiation period expenses, if consequential on a general average act, would have fallen within the exclusion in Rule C of loss sustained through delay, but I do not accept that it follows that they must therefore fall outside Rule F. Rule C applies to expenses and other sums claimed by way of general average as consequences of a general average act (as defined by Rule A). It does not apply to expenses covered by Rule F, which is concerned with sums which are expended or lost in mitigating or avoiding the sums which would otherwise be claimable as general average. By definition, sums recoverable under Rule F are not themselves allowable in general average, but are alternatives to sums which would be allowable. One can understand why, as a matter of policy, demurrage and similar indirect liabilities are not recoverable as general average, but it does not follow that such indirect liabilities should be irrecoverable if they are expended in order to mitigate what would otherwise be a larger general average claim.

38.              As for the cargo interests’ reliance on Rule XI, I find it hard to see why the fact that vessel-operating expenses are specifically allowed in one specific type of case, means that it should be presumed that they are excluded from every other type of case. In any event, the Rules start by saying that the lettered Rules apply save where the numbered Rules apply, and that makes it particularly difficult to justify the notion that a specific allowance in a numbered Rule should impliedly rule out such an allowance in a lettered Rule. Indeed, I understood the cargo interests in this case to accept that vessel-operating costs would be recoverable in a case where Rule F did apply (subject to their Rule C argument considered in para 37 above), and that seems to be consistent with what is said in the two books on general average to which I have referred.

39.              For these reasons, I would allow this appeal and restore the decision of the deputy judge.

40.              I agree with the judgment of Lord Neuberger.

41.              The York-Antwerp Rules have a status in shipping law similar to that of the Uniform Customs and Practices in the law relating to documentary credits. They depend wholly on contractual incorporation for their binding force. But they are designed to create a body of principle applicable internationally in a uniform way, although incorporated in shipping agreements of different kinds, governed by different laws. It will therefore rarely if ever be appropriate to imply matter into them which is not apparent from the natural meaning of the words, unless the implication is necessary to make them workable or intelligible or to avoid absurdity. Rule F is simplicity itself. It provides for the allowance of expenditure which is not allowable as general average expenditure but has successfully mitigated expenditure or sacrifice which would have been allowable as general average. The cost of maintaining the ship and crew during a period of delay which would not have occurred but for the peril but was necessary to enable the ransom to be reduced, is deemed to be general average up to the amount of the reduction.

42.              I appreciate that the practice of most average adjusters has been to disallow such expenditure. In the absence of a comprehensive body of case law (general average rarely reaches the courts), adjusters have adopted a variety of practices or rules of thumb to supplement the Rules. This is perhaps inevitable, but such practices are not law and there is a tendency in this field for them to lose sight of the basic concepts expressed in the Rules themselves. I suspect that this particular practice has been influenced by the second paragraph of Rule C and the limited scope of application of Rule XI. But the second paragraph of Rule C serves to limit the permissible heads of general average expenditure so as to exclude delay. There is no textual, indeed no rational reason why it should be taken to limit the permissible heads of expenditure which although not general average expenditure successfully mitigates something else that is. As for Rule XI, like the other numbered rules, that is a specific rule relating to ships entering a port or place of refuge, which does not impinge upon the general principles set out in the lettered rules, as applied to other situations.

43.              In my opinion, the appeal should be allowed.

44.              I agree that the appeal should be allowed for the reasons given by Lord Neuberger and Lord Sumption.

45.              Although a general average case was the origin of the English Commercial Court, general average cases are few and far between. The correct resolution of the present case has divided both general average practitioners and the courts, and the number of issues raised has tended to multiply as the case has progressed.

46.              The core question is simple. Where a vessel with its cargo has been seized by pirates, and the owners over a period succeed in negotiating down an initial ransom demand, can the owners include in general average not merely the ransom payment ultimately made, but also vessel and crew costs totalling US$160,213.95 incurred during the period of negotiation (“the negotiation period expenses”). The vessel MV Longchamp was boarded by pirates at 06.40 hours on 29 January 2009. A ransom demand of US$6m was made by the pirates at 14.05 on 30 January 2009 and was rejected by the owners as “too high” on 31 January 2009. The vessel had by then been taken to a position off Eyl on the coast of Somalia. Thereafter, negotiations took place lasting until 18.25 on 22 March 2009, when the pirates accepted the owners’ last offer of US$1.85m. The ransom was dropped at sea off Eyl on 27 March and the vessel was released to proceed on her voyage at 07.36 on 28 March 2009. The negotiation was in practice conducted by the owners. The negotiation period to which the relevant expenses relate runs from 14.05 on 30 January to 08.25 on 22 March 2009.

47.              The relevant bill of lading provided for any general average to be adjusted according to the York-Antwerp Rules 1974. The ransom payment and the costs of specialist negotiators are accepted as direct general average costs, falling within Rule A of those Rules. The question is whether the negotiation period expenses fall to be included in general average under Rule F, reading:

48.              The owners’ case is that: (i) if they had paid the US$6m ransom initially demanded, that “would have been allowable as general average” within the meaning of Rule F; (ii) instead of doing this, they entered into successful negotiations; (iii) the resulting reduction in the ransom payable from US$6m to US$1.85m avoided general average expense of US$4.15m; and (iv) the negotiation period expenses totalling US$160,213.95 can and should be treated as an extra expense incurred in place of the general average expense of US$4.15m avoided.

49.              Before the deputy judge, Mr Stephen Hofmeyr QC [2015] 1 Lloyd’s Rep 76, the focus was on point (i), whether the US$6m would, if paid, have been allowable as general average. He held that it would have been. The Court of Appeal agreed with the judge on this point. But it focused on a different aspect, whether Rule F was in principle applicable when all that had occurred was negotiation, in the event a long, rather than a short negotiation, but not an alternative course of action by which expenditure different in kind was incurred. In the Court of Appeal’s view, this was not a case of extra expense being incurred “in place of” another expense, but a case where the owners had no “real choice” or “true alternative” to pursue (per Hamblen LJ [2016] Bus LR 1285, paras 51-53) and there was “only one course of action open … namely to treat with the pirates … however long that might take” (per Sir Timothy Lloyd, para 99).

50.              I have considerable sympathy with the Court of Appeal’s instinct that Rule F was not designed with the present situation in mind. The classic circumstances in which it is treated as applying are cases where there is one obvious or natural course of action open to the owners following a general average event, but there is also some different action, which might if taken lead to a more generally beneficial outcome overall. This, it is common ground that Rule F in the 1974 Rules embraces situations where, instead of undergoing repairs in a port of refuge, the vessel is towed as is to destination, or the cargo is forwarded on another vessel, or, where, instead of discharging cargo in order to undertake dry-docking and repair, extra equipment is obtained to enable dry-docking and repair without such discharging. In the Canadian case of Western Canada Steamship Co Ltd v Canadian Commercial Corp [1960] 2 Lloyd’s Rep 313, decided in an era when sea transport was the norm, the cost of airfreighting a new propeller shaft from Wales to Singapore was allowed on the same basis. However, it is also clear that Rule F can apply in situations where the general average expense avoided would have been allowable under one of the numbered rules, setting out various specific situations in which various types of vessel or crewing costs are in principle recoverable as general average.

51.              In short, the focus of Rule F seems to me to have been correctly identified by Hoffmann LJ in Marida Ltd v Oswal Steel (The Bijela) [1993] 1 Lloyd’s Rep 411, 421, when he quoted with approval the then most recent edition of Lowndes & Rudolf on General Average and the York-Antwerp Rules (11th ed (1990), p 144), as saying that the substituted expenses must be:

52.              The cargo interests conceded in their written case, and I am prepared for present purposes to accept, that this does not exclude all possibility that Rule F might cover a situation in which, by some unusual or non-standard step, the owners are able to replace or reduce in amount an expense of one kind by incurring a lesser expense of the same kind. It does not however seem to me necessary to go even that far in the present case, since what is submitted is that the owners have, by incurring negotiation period expenses, consisting of vessel and crew costs, avoided a different kind of expense, namely extra ransom costs.

53.              What is however clear on any view is that Rule F is not intended to cover general average situations in which owners simply do what would in the ordinary course be expected of them in the interests of the common adventure. Where this is the position, the expenses incurred will be admissible, potentially, as “the direct consequence of the general average act”, allowable as general average under the first paragraph of Rule C, unless they are excluded as loss or damage incurred through delay under the second paragraph of Rule C. One qualification to this is however significant. Even if expenses would otherwise be excluded as loss or damage incurred through delay, they may nevertheless be admitted if they fall within one of the numbered heads. Rule XI(a) in particular covers costs of crew and vessel’s fuel and stores during the prolongation of a voyage occasioned by a vessel entering a port or place of refuge or returning to her loading port or place in circumstances falling within Rule X(a). Rule XI(b) covers, broadly, crew costs when a vessel has entered or been detained in any port of place in consequence of, or for repairs following, a general average event. It is unnecessary to examine the precise ambit of these provisions. What matters is that they constitute very specific qualifications of the exclusion by the second paragraph of Rule C of loss or damage through delay. There is no equivalent qualification which could cover the present case. Hence the owners’ reliance on Rule F.

54.              In order to bring the case within Rule F, the owners have to show - and the onus is on them both as a matter of general principle and specifically under Rule E - that they incurred “extra expense”, that this was in place of another expense and that that other expense would have been allowable as general average; and, once they have done that, the extra expense is only allowable up to the amount of the general average expense avoided. I have little difficulty with accepting the negotiation period expenses as an extra expense in the sense that they were over and above anything that would, taking Mr Hofmeyr QC’s words ([2015] 1 Lloyd’s Rep 76, para 89), “ordinarily have been incurred” on such a voyage. I have already covered what may have been contemplated in Rule F by the use of the phrase “in place of another expense”.

55.              Rule F only applies if the other expense “would have been allowable as general average” and subject to the condition that the extra or substituted expense is allowable “only up to the amount of the general average expense avoided” by not incurring that other expense. Viewing Rule F as a whole, it is clear that the owners must show that, had they incurred the other expense, the costs it would have involved could validly have been treated as general average - their right to include extra or substituted expenses as general average being limited under Rule F by the extent to which the hypothetical other expense could have been so treated. The last sentence of para 19 of Lord Neuberger’s judgment accepts this. It seems therefore of only academic interest to debate whether the same conclusion would flow from the words “would have been allowable as general average” by themselves. For my part, however, I consider that it would have done. The word “allowable” must take one back to the requirements of Rule A as well as Rule C. Rule A apart, there would be no indication what is meant by “general average” or an expense allowable in general average. The onus placed under Rule E on any party claiming in general average to show that the loss or expense claimed is “properly allowable as general average” must carry one back to Rule A as well as C. In this connection, it is also notable that the French word used throughout Rules C, E and F (admis, admise, admission or admissible) has in each case the same root.

56.              The “Hudson conundrum”, to which Lord Neuberger refers (para 18 above), does not lead to any different conclusion. In most, if not all, circumstances in which Rule F applies, there will be a prima facie or standard course of action to be taken in the face of the general average event, but the owners will, by adopting some unusual means, have arrived at an alternative solution to further the common adventure. In such cases, the other paying parties will be hard pressed to suggest that the prima facie or standard reaction would have been unreasonable. There is a parallel here with the duty to mitigate, which is not lightly to be imposed or treated as broken. The Court of Appeal ([2016] Bus LR 1285, paras 73-74) was in this respect right in my view to agree with the deputy judge’s general conclusion of principle ([2015] 1 Lloyd’s Rep 73, para 77) that the hypothetical other expense must be one which would have been reasonably incurred in a sense “interpreted and applied with a sufficient degree of latitude to give rule F practical effect”. In many cases, the differences between the two courses (the standard and that adopted) may not be large, and both may easily be reconciled as reasonable reactions. Rule F is also careful, by its concluding words, to recognise that the extra expense may not be less than that which would have resulted from taking the standard course. In such a case, Rule F performs the valuable function of allowing recovery up to the amount which would have been recoverable had the standard course been adopted.

57.              Turning specifically to an unusual situation like the present: if there is no course at all open to take, the expenses of which would have been allowable as general average, then matters must run their course. If a ransom is demanded and paid in an amount which is unreasonable to pay, the only amount allowable in general average will be whatever lesser amount it would have been reasonable, after negotiation, to pay. If the negotiation period expenses are regarded as an extra expense incurred in place of the amount of the ransom avoided by the negotiation, they can be recoverable at most only so far as the negotiation avoided the making of a ransom payment which it would have been reasonable to pay.

58.              On no view is there any basis for reading into the clear language of Rule F an entirely artificial assumption that, when judging whether the other expense would have been allowable as general average, the possibility of incurring the extra expense in place of that other expense must be ignored. To do so would be flatly contrary to the language and evident intent of Rule F. The reasoning and decision of the House of Lords in Marida Ltd v Oswal Steel (The Bijela) [1994] 1 WLR 615 turned on the very different wording of Rules X(b) and XIV, under which the express assumption in Rule XIV that “such [temporary] repairs had not been effected there” could be, and fell to be, read into Rule X(b) when considering what repair expenses would have been necessary and allowed in general average for the purposes of Rules X(b) and XIV, read together.

59.              Both the courts below have in this case concluded that it would have been reasonable for the owners to accept and pay the first ransom demand of US$6m. The Court of Appeal has however decided the case against the owners on the basis that they faced no real choice but to negotiate, however long the negotiation might take. There is to my mind a tension between the two strands of the Court of Appeal’s reasoning. If it would have been reasonable to accept and pay the first demand, then the owners were on the face of it taking a stand by seeking an even more reasonable deal in the interests of all concerned in the common adventure. It is not apparent that Rule F could not extend to such a course, if, as here, it involved the owners in some expense in the form of additional crew and vessel expenses.

60.              But I find even more difficult the joint conclusion of both courts below that it would have been reasonable for the owners to meet the first ransom demand. The deputy judge found difficulty in seeing how any ransom payment could be described as reasonable: [2015] 1 Lloyd’s Rep 76, para 98. He said:

61.              However, this was as far as the deputy judge was prepared to go. Leaving aside exceptional circumstances, where the value of the ransom demanded clearly exceeded the value of the property involved in the venture, he thought it obvious that it would not be reasonable to say that an owner under an obligation to proceed with due despatch had not reasonably incurred a ransom paid. He went on, at para 99:

“Even if it may be said that, by January 2009, a pattern of dealing between Somali pirates and shipowners had developed, as described by David Steel J in Masefield AG v Amlin Corporate Member Ltd (The Bunga Melati Dua) [2010] 2 All ER 593’ [2010] 1 Lloyd's Rep 509 at paras 19, 23, 25 and 26 (affirmed on appeal: [2011] 1 Lloyd’s Rep 630; [2011] 1 WLR 2012), such a pattern would not remove the potential for unreasonable, irrational and illogical behaviour.”

In support, the deputy judge said that negotiation was an uncertain process and it “was not possible to state with reasonable certainty when the ransom demand was made that the amount of the ransom would inevitably be significantly reduced by the process of negotiation” (para 100). Whether or not it is “possible to state with reasonable certainty” that a negotiation will achieve significant success cannot however be the test of whether or not negotiation should reasonably be essayed. The deputy judge also derived comfort from his conclusion on this point from the consideration that “natural justice requires that all should contribute to the substituted expenses incurred” (para 103). A difficulty about this observation is that nothing in Rule F could enable cargo interests to recover any matching loss or damage that they might suffer from the delay during negotiations. The cargo was in fact perishable, even though in the event it survived the rigours it underwent without apparent deterioration. As the Court of Appeal correctly recognised in this connection (para 51), whether or not an item falls within general average depends on the proper interpretation of the York-Antwerp Rules. They represent a balanced framework, negotiated over time between all interests involved.

62.              As I have already observed, the Court of Appeal’s reasoning involves a potential tension between the approaches taken to the scope of Rule F and to the issue of the reasonableness of paying the initial ransom demand. In the former context, Hamblen LJ said this [2016] Bus LR 1285, paras 43-46:

“43.     Some support for the cargo interests’ approach is to be found in the evidence. Thus, it does not appear that the owners ever considered that they faced a choice. The owners’ crisis management and negotiation team were set up before any ransom demand had been made. From the outset the goal was to negotiate to obtain release of the vessel upon payment of a ransom, but in a reduced amount. There is no evidence to suggest that they ever considered choosing between paying the ransom on demand and paying a lesser sum following negotiation.

44.       This is also borne out by the advisory committee’s stated experience, which is that in all Somali piracy cases the same course of action is taken, namely to negotiate and pay a reduced ransom leading to release of the vessel. Again it does not appear that there is considered to be a choice of payment on demand.

45.       In my judgment this failure to recognise that there is a choice reflects the reality, which is that payment on demand is simply a different way of going about the same course of action and not a true alternative course of action. Whether or not the ransom is paid on demand there will still be a negotiation, there will still be delay, there will still be the incurring of vessel and crew running costs during the period of delay. In either case the same expenses will be incurred; the difference is only in their extent.

46.       In this case, for example, there was a period of delay between the hijacking and the first ransom demand. Even if that first demand had been accepted, it does not follow that it would have been agreed. As the majority of the advisory committee state, the unprecedented acceptance of the ransom on demand may well have been ‘met by a demand from the pirates for a still higher figure’. Even if that was not the case, it would still have been necessary to negotiate and agree matters relating to place and method of payment and to the release of vessel and crew. Thus in this case it is to be noted that there was a period of six days between the agreement of the ransom and the release of the vessel.”

63.              In the latter context, however, Hamblen LJ said, at paras 77-84:

81.       In my judgment, if, as stated in the Masefield case [2010] 2 All ER 593, ‘the safest, most timely and effective means to secure the release’ of a ship and crew was to pay a ransom, it follows that the most safe, timely and effective means of so doing is to pay as soon as possible. It may be that the general practice was to try to negotiate the ransom down, but that does not mean that it would be unreasonable to pay the ransom straight away so as to avert the very real danger to vessel, cargo and crew as quickly and effectively as possible. Nor can a course of action which procures such real and tangible benefits be regarded as an ‘artificial invention’.

82.       Further, in my judgment the reasons given by the judge are all cogent and compelling reasons for concluding that payment of the initial ransom sum would have been reasonable.

83.       Further reasons for supporting that conclusion include the following:

(1)       The effect of the delay involved in seeking to negotiate a lower ransom is to keep the vessel, cargo and crew in peril, with all the risks of saying ‘no’ to pirates, who are violent, armed criminals.

(2)       The vessel and cargo were under the control of the pirates. As such, there were obvious dangers should there be a storm or other peril of the sea.

(3)       The owners knew that there had been a firefight during the capture of the vessel and that a crew member had been wounded.

(4)       Although, as matters turned out, the pirates’ main negotiator was said to be a ‘calm, rational communicator’ who never resorted to threats or other coercive tactics, the owners had no reason to assume that.

(5)       This was just one of many ‘known unknowns’ facing the owners.

84.       For all these reasons I conclude that it cannot be shown that the judge was wrong to find that payment of the initial ransom demand would have been reasonable. It follows that I would dismiss the appeal on this issue.”

I note that, after the quotation from the Masefield case [2010] 2 All ER 593 in para 81 of Hamblen LJ’s judgment, the words “to pay a ransom” are not a correct citation. The actual words in the Masefield judgment were “to negotiate and subsequently pay a ransom”.

64.              The cargo interests rely on the apparent acceptance both by the deputy judge ([2015] 1 Lloyd’s Rep 76, para 99, quoted in para 61 above) and by the Court of Appeal ([2016] Bus LR 1285, paras 43, 44, 46 and 81) of a general practice to negotiate any ransom demand down over a period. The owners object that there is no evidence justifying any such conclusion. But their own skeleton argument for the trial stated that:

“This is not to say that paying the first-demanded ransom is ever likely in fact to be a reasonable course of action. In reality, where there is the option of entering into negotiations with pirates, it will almost always be the right thing to do. But when considering the allowability in GA of the costs of the negotiation, the relevant alternative scenario to be considered is the one in which there is no negotiation possible/available.”

The last sentence reflects the owners’ then case, which relied on a suggested analogy with the reasoning of the House of Lords in The Bijela, discussed in para 58 above.

65.              The trial in the present case was a Commercial Court trial on the documents. It seems clear that the deputy judge treated himself as entitled to rely on all the material before him when considering the factual position, including the statements in the Masefield case and the Report of the Advisory Committee of the Association of Average Adjusters on the present case, which he summarised at some length, while recording that it was common ground that it was not binding on the court. I see no reason to regard either the deputy judge or the Court of Appeal as having erred in this respect. David Steel J’s judgment in the Masefield case noted ([2010] 2 All ER 593, para 14) that the initial ransom demand in that case of US$2m in August 2008

“was all of a piece with the process of Somali hijacking. Fortunately the process of negotiating such a demand and making an agreed payment had invariably led to the release of all vessels involved. Against that background, I did not understand it to be controversial that the actual prospects of recovery of the cargo as at 18 September 2008 were good.”

66.              Other relevant factors on the issue of reasonableness are that the range of potential values as assessed at the time was between US$5m and US$7m (with US$5.4m being later established as the correct figure). A demand of US$6m self-evidently exhausted or very nearly exhausted all interests involved. Further, any indication of agreement to pay anything like the initial demand would almost inevitably have fed a suspicion on the pirates’ part that they had demanded far too little, and would have complicated matters then and for the future. In contrast, and in the light of the past experience of other shipowners whose vessels had been seized by Somalian pirates, there was on the face of it every reason to give effect to what was evidently the present owners’ immediate reaction, that is to hire experienced negotiators and engage on a time-consuming and painstaking process of negotiation.

67.              In reaching a conclusion that it would have been reasonable for the owners to capitulate in response to the very first demand, the courts below were making an evaluative judgment on the basis of documentary evidence and material. This is not a situation in which their evaluation commands a large inherent advantage, compared with that which the Supreme Court is in a position to make, although of course it merits weight and it is for the cargo interests to show that it was wrong. In the light of all the circumstances, the cargo interests have satisfied me that it was wrong. I am unable to accept the evaluative judgment reached by both courts below to the effect that it would have been reasonable for the owners to pay the initial ransom, and that, had they done so, they could have required the full US$6m to be treated as general average. The reasoning of the courts below appears to me contrary to all the relevant indications as to how the owners actually acted and would have been expected to act. It is clear that the owners never contemplated the sort of remarkable capitulation that payment of the initial ransom would have involved, and that it would have taken them and other shipowners into uncharted territory, as opposed to a relatively familiar negotiation process, had they ever done so. They would in my opinion clearly have been acting unreasonably in the circumstances had they done so.

68.              The case has been fought and decided on the basis that this is the critical issue. No alternative case has been advanced to the effect that negotiations would or might have led to a settlement at some lesser figure which might have covered some, no doubt lesser figure of negotiation period expenses. Lord Neuberger notes (para 20 above) that the actual negotiation period expenses claimed of US$160,000 would have been covered (together with the negotiators’ expenses) by a ransom payment of around US$2.4m. That is mathematically correct. But it does not reflect the reality which has to be addressed. It postulates immediate agreement on or about 30 January 2009 on a ransom of US$2.4m. Yet, even in early March 2009 the pirates were still looking for a ransom of US$3m, reduced on 2 March 2009 to US$2m. Depending on where one dates and places a reasonable settlement at a reasonable settlement figure, it is clear that, on this basis, any recoverable negotiation period expenses would be considerably reduced below the US$160,000-odd claimed. The case has not been put on a basis which required or allows now for any such hypothetical exercise (of assessing when and at what figure below US$6m a reasonable settlement could have been achieved) to be undertaken. The owners have established that Rule F is in principle capable of applying to negotiation period expenses, which may well be the principle which this litigation is about. But I do not think that they have established on the facts that they have any claim on the only factual basis on which the case has been put. I would therefore dismiss this appeal, albeit for reasons different from those given by the Court of Appeal.