Petroleo Brasileiro S.A. v E.N.E. Kos 1 Limited

This appeal is about the rights of the owner of a time-chartered ship after the ship has been lawfully withdrawn for non-payment of hire. The question must often have arisen in practice but, oddly enough, there is no direct authority upon it.

The MT Kos is a 301,000 mt VLCC. She was time chartered by her owners to Petroleo Brasileiro SA on 2 June 2006 for 36 months plus or minus 15 days at charterers’ option. The charterparty, which was on the Shelltime 3 Form, contained a standard form of withdrawal clause providing that if hire was not paid when due, the owners should have the right to withdraw the vessel “without prejudice to any claim owners may otherwise have on charterers under this charter.” It is increasingly common for such contracts to include anti-technicality clauses requiring notice to be given before this right is exercised. But for whatever reason no anti-technicality clause was included in this case. So when, on 31 May 2008, charterers failed to make the advance payment required for the month of June, the owners were entitled to withdraw the MT Kos, and did so at 14.41 GMT on 2 June 2008. It is agreed between the parties to the appeal that the charterers’ failure to pay hire was not a repudiatory breach of contract.

At the time of the withdrawal, the MT Kos was at Angra dos Reis in Brazil, where she had just completed the loading of a parcel of cargo for the charterers’ account in accordance with their orders. She was awaiting a second parcel, which in the event was not loaded. There were no bills of lading in the hands of third parties. On 2 and 3 June there was a number of exchanges between the parties. The charterers tried to persuade the owners to cancel the withdrawal. The owners refused. Their position was set out in a message at 11.30 GMT on 3 June 2008. They said that they would be willing to reinstate the charterparty or continue on a voyage basis, but only at the current market rate, which was much higher than the charterparty rate. Otherwise, they required the charterers to make prompt arrangements to receive back their cargo. Ultimately, after further fruitless exchanges, the charterers told the owners at 21.36 GMT on 3 June 2008 that they would arrange for the terminal to receive back the cargo. The arrangements were duly made, and discharge of the cargo was completed at 06.00 GMT on 5 June 2008. It is agreed that if the charterers had begun to make arrangements for the discharge of their cargo as soon as they received the owners’ notice of withdrawal, the vessel would have been detained at Angra dos Reis for one day. As it was, she was detained there for 2.64 days.

The issue before us is whether the owners are entitled to be paid for the service of the vessel during that 2.64 days, and for bunkers consumed in the same period. Leaving to one side points which have fallen by the wayside at earlier stages of these proceedings, their claim is put forward on four bases: (i) under clause 13 of the charterparty; (ii) under an express or implied new contract made after the vessel was withdrawn, to pay for the time and bunkers; (iii) on the ground of unjust enrichment; and (iv) under the law of bailment. The judge, Andrew Smith J, held that they were entitled to succeed on basis (iv), but rejected every other basis which they put forward. The Court of Appeal (Longmore and Smith LJJ and Sir Mark Waller) rejected the claim on all four bases, except that they allowed the owners to recover the value of bunkers consumed in actually discharging the cargo.

The argument that there was a new contract turns entirely on the facts and can be shortly dealt with. Once the charterparty came to an end, the owners no longer had any obligation to carry the cargo to its destination or to discharge it. Their duty was to make it available to the charterers. It was then for the charterers to make any necessary arrangements for discharge. For a day and a half after the notice of withdrawal, they did nothing because they declined to accept that the owners were entitled to withdraw the vessel. Each party was trying to persuade the other to resume the contractual service (or a variant of it) on its own terms. Each of them rejected the other’s terms. On the footing that the owners were not willing to treat the old contract as subsisting and that no agreement could be reached upon a new one, both parties then submitted to the inevitable. Owners called on charterers to take delivery of their cargo, as charterers in any event were bound to do. The charterers then got on with it. Both courts below held that it was impossible to spell a new contract out of these facts. I agree.

Under all the remaining heads of claim, the charterers’ argument is substantially the same, namely that any delay or loss arising from the need to discharge the cargo results from the owners’ decision to withdraw. That was a decision made at their own election and for their own commercial purposes. The owners, it is said, must bear the adverse as well as the beneficial consequences of an optional decision made in their own interest. It is clear that this consideration influenced both courts below, and that it was decisive in the minds of the Court of Appeal.

The factual premise of the argument is of course correct. It is axiomatic that a withdrawal clause operates at the election of owners, and not automatically. Two main consequences follow from this. The first is that owners will not exercise their right of withdrawal unless it is in their commercial interest to do so. Usually, this will be because market rates of hire have risen. But it may be in owners’ interest to withdraw the vessel even if they have not risen, for example, where the charterers are insolvent or owners depend on prompt payment to fund payments under a head charter or charterers’ payment record occasions administrative or other difficulties. The second consequence is that any failure on the part of the charterers to pay hire when it falls due will not of itself entitle the owners to damages representing the loss of the bargain or the expenses of termination simply because the owners respond by withdrawing the vessel. This is because the non-payment does not itself destroy the bargain or occasion the expenses, unless in the circumstances it is a repudiation which owners have accepted as such. But the present claim is not a claim for damages, and the non-payment of the June 2008 hire payment in this case was not a repudiation. This, however, is as much as can usefully be said. The fact that rather than perform the contract the owners found it more advantageous to exercise an express right of termination is morally and legally neutral. There are no standards by which the owners’ reasons may be judged, other than those to be found in the contract. There is no legal policy specific to termination rights restricting their availability or the consequences of their exercise more narrowly than does the language of the contract or the general law. More generally, the reasons for any particular withdrawal cannot affect the principle to be applied in resolving an issue like the present one.

Clause 13 provides, so far as relevant:

“The master (although appointed by owners) shall be under the orders and direction of charterers as regards employment of the vessel, agency or other arrangements. Bill[s] of lading are to be signed as charterers or their agents may direct, without prejudice to this charter... charterers hereby indemnify owners against all consequences or liabilities that may arise from the master, charterers or their agents signing bills of lading or other documents, or from the master otherwise complying with charterers’ or their agents’ orders…”

Clause 13 is the employment and indemnity clause which is found in most modern forms of time charter. The indemnity reflects the breadth of the powers conferred on the charterers as to the employment of the vessel. As Devlin J observed in Royal Greek Government v Minister of Transport (1949) 83 Ll L Rep 228, 234, “if [the owner] is to surrender his freedom of choice and put his master under the orders of the charterer, there is nothing unreasonable in his stipulating for a complete indemnity in return.” Indeed, the courts have held that, subject to the express terms of any particular charterparty and to the limitations which I shall consider below, the indemnity is not just ‘not unreasonable’. It is necessary. It will generally be implied even in forms of time charter (such as the New York Produce Exchange Form) where it is not expressed.

The scope of the indemnity in clause 13, like that of the corresponding implied term, is very wide (“all consequences or liabilities that may arise”). But it is not “complete”, nor is it unlimited.

In the first place, it has to be read in the context of the owners’ obligations under the charterparty as a whole. The owners are not entitled to an indemnity against things for which they are being remunerated by the payment of hire. There is therefore no indemnity in respect of the ordinary risks and costs associated with the performance of the chartered service. The purpose of the indemnity is to protect them against losses arising from risks or costs which they have not expressly or implicitly agreed in the charterparty to bear. What risks or costs the owners have agreed to bear may depend on the construction of other relevant provisions of the contract, or on an informed judgment of the broad range of physical and commercial hazards which are normally incidental to the chartered service, or on some combination of the two. The classic example of a loss within the indemnity, and probably the commonest in practice, is one which arises from the master complying with the charterers’ direction to sign bills of lading on terms of carriage more onerous than those of the charterparty. But the indemnity has been held to be applicable in principle to a wide variety of other circumstances, including compliance with an order to load cargo which is dangerous even on the footing that appropriate care is taken of it, or an order to proceed to a legally unsafe port. On the other hand, the indemnity will not apply to risks which the owners have contractually assumed, which will usually be the case where they arise from, for example, their own negligence or breach of contract or consequences such as marine fouling which are incidental to the service for which the vessel was required to be available.

Secondly, clause 13 itself limits the indemnity to losses which were caused by complying with the charterers’ orders. Like all questions of causation, this one is sensitive to the legal context in which it arises. It depends on the intended scope of the indemnity as a matter of construction, which is necessarily informed by its purpose. We are not therefore concerned with questions of remoteness and foreseeability of the kind which would arise in the law of damages, where the object is to limit the range of consequences for which a wrongdoer may be said to have assumed responsibility in the eyes of the law. Indeed, as Sir Donald Nicholls V-C pointed out in Triad Shipping Co v Stellar Chartering & Brokerage Inc (The Island Archon) [1994] 2 Lloyd’s Rep 227, 238, the more foreseeable the owners’ loss, the more likely it is to be an ordinary incident of the chartered service and therefore outside the scope of the indemnity. The real question is whether the charterers’ order was an effective cause of the owner having to bear a risk or cost of a kind which he had not contractually agreed to bear. I use the expression “effective cause” in contrast to a mere “but for” cause which does no more than provide the occasion for some other factor unrelated to the charterers’ order to operate. If the charterers’ order was an effective cause in this sense, it does not matter whether it was the only one.

For present purposes, the relevant order of the charterers was the order to load the parcel of cargo which was on board the vessel when it was withdrawn. In my judgment the loss claimed by owners was the consequence of that order. The need to discharge the cargo in the owners’ time arose from the combination of two factors, namely (i) that the cargo had been loaded, and (ii) that the purpose for which it had been loaded (ie carriage under the charterparty to its destination) had come to an end with the termination of the charterparty. In other words, the cargo which charterers had ordered the vessel to load was still on board when the charterparty came to an end. On any realistic view, this was because the charterers had put it there. The analysis would have been exactly the same if the charterparty had come to an end for any other reason with cargo still on board, for example by frustration or expiry at the end of the contractual term.

Andrew Smith J and the Court of Appeal both rejected the claim under clause 13 on the ground that the true cause was the owners’ withdrawal of the vessel. The judge said at para 35 of his judgment that the owners’ claims were “too remote” from the order to load. Longmore LJ, giving the reasons of the Court of Appeal, observed at para 15 that it was

“not a natural consequence of ordering [the cargo] to be loaded that it would have to be discharged at the self-same port. The true cause of the necessity for the discharge of the cargo was the fact that, in the light of the withdrawal, the owners required the charterers to discharge the cargo.”

In effect, therefore, both courts below found that the withdrawal of the vessel was an independent cause of the loss, breaking the chain of causation between the order to load the cargo and the detention of the vessel after withdrawal.

The difficulty about this is that because the cargo had been loaded, it had to be discharged somewhere, if not at the port of loading then at its destination or possibly at an intermediate port. The owners’ decision to withdraw the vessel or, to be precise, the adventitious timing of that decision, merely determined the place at which the discharge of the cargo occurred. If the owners were to withdraw the vessel, they had to do it promptly upon hire going into default, and it so happened this was when the vessel was still at the port of loading. But the precise timing of the withdrawal and location of the discharge are irrelevant to the owners’ loss. If the vessel had been withdrawn immediately before discharge at the destination, the consequence would have been exactly the same. It is of course true that discharge at the destination would have been a great deal more beneficial to the charterers than discharge at the port of loading. This is a point that seems to have influenced the Court of Appeal. But a claim under clause 13 does not depend on the benefit conferred on the charterers. It depends on the detriment to the owners. They would have suffered a detriment of much the same kind wherever the vessel had discharged.

It is fair to say that it was only because of the withdrawal of the vessel that the subsequent discharge of the cargo at Angra dos Reis had to be done in the owners’ time and without earning contractual hire. But that is the very reason why the detention of the vessel falls within the indemnity. The need to discharge the cargo in their own time and at their own expense was not an ordinary incident of the chartered service and was not a risk that the owners assumed under the contract. It arose after the chartered service had come to an end in accordance with the withdrawal clause in the contract. Mr Baker QC for the charterers asked rhetorically whether, in that case, the owners would be entitled to claim the cost of sending the vessel in ballast from Angra dos Reis to somewhere else where she could start employment under a new charterparty. But a claim like that, although ultimately dependent on its particular facts, would be likely to fall on the other side of the line. The need for a ballast voyage before a vessel can begin her next employment is an ordinary commercial risk associated with the trading of the vessel under a time charter.

It remains to consider the measure of the indemnity on the facts of this case:

Strictly speaking, this makes it unnecessary to address any of the other legal bases put forward by the owners in support of their claim. But I propose to deal with the question whether the owners were also entitled to succeed at common law as non-contractual bailees of the cargo after the withdrawal of the vessel. I do so partly out of respect for the trial judge who decided the case on that basis, and partly because I think that the commercial and legal logic of the claim in bailment is close to the logic which brings it within clause 13 and would bring it within any corresponding implied term. On the whole, one would expect a coherent system of law to produce a consistent answer under both heads, and in my judgment it does.

Unlike many civil law systems, English law does not allow a general right of recovery for benefits conferred on others or expenses incurred in the course of conferring them. In the pejorative phrase which has become habitual, there is no recovery for benefits “officiously” conferred. In Falcke v Scottish Imperial Insurance Co (1886) 34 Ch D 234, 248 Bowen LJ said:

“The general principle is, beyond all question, that work and labour done or money expended by one man to preserve or benefit the property of another do not according to English law create any lien upon the property saved or benefited, nor, even if standing alone, create any obligation to repay the expenditure. Liabilities are not to be forced upon people behind their backs any more than you can confer a benefit upon a man against his will.”

While this remains the general principle, the exceptions have over the years become more important than the rule. The particular feature of the present case which makes it difficult to apply the general rule is that the original bailment of the cargo had occurred under a previous contractual relationship. The bailment was therefore consensual, albeit that after the withdrawal of the MT Kos from the time charter, it was no longer contractual. It is common ground, and clear on the authorities that in these circumstances, the owners had a continuing duty to take reasonable care of the cargo, which they could not escape except by retaining it until arrangements were made to discharge it. But the owners had in no sense officiously put themselves in this position, nor had they (as the charterers put it in argument), “voluntarily assumed” possession of the goods. There is a thin, but consistent line of authority which deals with the legal consequences of this situation.

In Gaudet v Brown (1873) LR 5 PC 134 (“Cargo ex Argos”), petroleum was shipped in London on the Argos under a bill of lading providing for delivery at Le Havre. The vessel arrived at Le Havre in the later stages of the Franco-Prussian war, when the port was full of munitions, and the landing of flammable cargoes was forbidden. The master therefore discharged the petroleum into lighters in the outer harbour, and it seems that the shippers (who had retained the bill of lading) could have taken delivery of it there and transported it elsewhere. But they failed to present the bill of lading or to make any arrangements to receive it. Having waited for as long as the port authorities would allow him to, the master reshipped the cargo and carried it back to London. The owners then successfully sued the shippers for freight for the return voyage. The case appears to have been decided on the footing that the contract of carriage was at an end when the Argos left Le Havre for London, either because the contractual service had been completed or because the contract was frustrated at Le Havre. The ground of the decision was expressed at pp 165-166 as follows:

“…not merely is a power given, but a duty is cast on the master in many cases of accident and emergency to act for the safety of the cargo, in such manner as may be best under the circumstances in which it may be placed; and that, as a correlative right, he is entitled to charge its owner with the expenses properly incurred in so doing… In a case like the present, where the goods could neither be landed nor remain where they were, it seems to be a legitimate extension of the implied agency of the master to hold that, in the absence of all advices, he had authority to carry or send them on to such other place as in his judgment, prudently exercised, appeared to be most convenient for their owner; and if so, it will follow from established principles that the expenses properly incurred may be charged to him… The authority of the master being founded on necessity would not have arisen if he could have obtained instructions from the defendant or his assignees. But under the circumstances this was not possible.”

A year later, the Court of Exchequer reached a very similar conclusion in Great Northern Railway Co v Swaffield (1874) LR 9 Ex 132. Mr Swaffield sent his horse by railway to a station at Sandy. The horse arrived late at night, and the railway company lodged the horse overnight for their own account at a livery stable. Mr Swaffield failed to collect it on the following morning. The only basis on which he was prepared to give any instructions about the fate of his horse was that the railway company assumed all responsibility for storing and delivering it to him from the time of its arrival at Sandy. After four months of this, the railway company lost patience. They unilaterally delivered the horse to Mr Swaffield’s farm and then sued him for the livery charges to date. As in Cargo ex Argos, the case was decided on the footing that the contract of carriage had come to an end, in this case on the day after the arrival of the horse at Sandy, when the performance required of them as carriers was completed. Counsel did not refer to Cargo ex Argos. But Baron Pollock drew attention to it in the course of argument and based his judgment upon it. Having referred to previous authority to the effect that the railway company was bound to take reasonable care of the horse notwithstanding the termination of the contract of carriage, he observed (p 138) that “if there were that duty without the correlative right, it would be a manifest injustice.” In his concurring judgment, at page 136, Kelly CB treated the principle as applying because it was necessary for the railway company to incur the expenditure. “They had no choice unless they would leave the horse at the station or in the high road to his own danger and the danger of other people.”

The principle applied in these cases has commonly been analysed as depending on the agency of necessity of the carrier, which indeed is how Longmore LJ analysed it in his judgment in this case. The existence of a coherent doctrine of agency of necessity has occasionally been doubted: see Goff & Jones, The Law of Unjust Enrichment, 8th ed (2011), para. 18-50, where it is suggested that “the cases in which it has been invoked are now best understood in other ways.” But so far as the doctrine does have a coherent existence, the case law requires that a bailee of goods should have taken steps in an emergency for the sole benefit of the cargo in circumstances where it was impossible to communicate with the owners of the goods. On that ground, the Court of Appeal held that the doctrine had no application to the present case, and that that was the end of the matter.

It is true that in Cargo ex Argos the Privy Council used the language of agency and necessity. But the master of the Argos was not in fact acting as the cargo-owner’s agent, as he would have been if (for example) he had purported to bind him to a contract with a third party, such as a lighterman or a warehouseman. On the face of it, he was simply carrying the goods back to London on behalf of his owners, in circumstances where there was no contract to do so but no reasonable or practical alternative. His claim was for additional remuneration for his own services, in excess of the performance required of him under the contract. As for Swaffield, although Baron Pollock cited authority on agency of necessity, on its facts there was no emergency and no agency. The true basis of the judgments in Swaffield was that where the property was originally bailed under a contract of carriage and the carrier had no choice but to remain in possession after the contract had ended, the existence of a continuing duty to care for the cargo was a sufficient basis for imposing on its owner an obligation to pay. Pollock B, I think rightly, regarded this as the principle on which Cargo ex Argos was really founded.

This view of the matter was accepted by the House of Lords in China Pacific SA v Food Corpn of India (The Winson) [1982] AC 939. The facts were that the Winson, bound for Bombay with a cargo of wheat, stranded on a reef in the South China Sea. Salvors retained on behalf of the ship and cargo interests off-loaded the wheat into barges and took it to Manila, where it was stored for their account in warehouses. It was common ground that storage under cover was necessary to prevent deterioration of the wheat, and that upon its arrival at the warehouse the salvage services came to an end. Some time after that, the owners gave notice that they were abandoning the voyage and the contract of carriage thereupon came to an end. The salvors wrote to the cargo-owners’ solicitors asking them to take delivery of their property, but received no answer. On these facts the salvors were bailees under the salvage agreement from the time that the cargo was taken off the stranded vessel until it reached the warehouse, and were thereafter non-contractual bailees until the cargo-owners finally took possession of the wheat from the warehouse. The issue was whether the cargo-owners were liable to the salvors for warehouse charges incurred up to the time when the owners gave notice that they abandoned the voyage. The cargo-owners accepted liability for the charges after that point, but contended that while the contract of carriage subsisted the warehouse charges were the responsibility of the carriers alone. It was held that the salvors were entitled to succeed. The leading speech was given by Lord Diplock, with whom the rest of the House agreed. Lord Diplock considered (p 957) that the case turned on the application of “well known and basic principles of the common law of salvage, of bailment and of lien.” He expressed the principle (at p 960) as follows:

“… the bailment which up to the conclusion of the salvage services had been a bailment for valuable consideration became a gratuitous bailment; and so long as that relationship of bailor and bailee continued to subsist the salvors, under the ordinary principles of the law of bailment too well known and too well-established to call for any citation of authority, owed a duty of care to the cargo owner to take such measures to preserve the salved wheat from deterioration by exposure to the elements as a man of ordinary prudence would take for the preservation of his own property. For any breach of such duty the bailee is liable to his bailor in damages for any diminution in value of the goods consequent upon his failure to take such measures; and if he fulfils that duty he has, in my view, a correlative right to charge the owner of the goods with the expenses reasonably incurred in doing so.”

He regarded this as being the principle applied in Cargo ex Argos and Swaffield in which the decisive facts, on his analysis, had been (i) that the bailee was left in possession of the goods after the termination of the contract under which the bailment had originally been made, and (ii) that in the absence of any contrary instructions from the cargo-owner, the warehousing of the goods was necessary for their preservation: see p 960G-H. Lord Diplock added, at p 961:

“It is, of course, true that in English law a mere stranger cannot compel an owner of goods to pay for a benefit bestowed upon him against his will; but this latter principle does not apply where there is a pre-existing legal relationship between the owner of the goods and the bestower of the benefit, such as that of bailor and bailee, which imposes upon the bestower of the benefit a legal duty of care in respect of the preservation of the goods that is owed by him to their owner.”

Lord Diplock went on to consider the doctrine of agency of necessity, because it had been submitted on behalf of the cargo-owners that a bailee in possession of goods could have authority on that basis only if it was impossible to communicate with the owner of the goods. The argument was that although the cargo-owners were not very communicative, they were never actually out of contact. This submission was rejected because the restrictions placed by the case law on a bailee’s authority as an agent of necessity applied only where the bailee was acting as a true agent, ie by purporting to bind the bailor to arrangements with third parties. They did not apply to a bailee’s right to reimbursement of his own expenses. It was sufficient for that purpose that the bailor should have failed to give instructions: see pp 961G-962B. It is clear that the relevance of this last point was that if the owner of the goods had given instructions, the salvor could by complying with them have relieved himself of any further responsibility. The decisive point, and the sense in which the word “necessity” is used in these cases, is that if the bailee is in a position where he has no way of discharging his responsibility to care for the goods without incurring loss or expense, then the loss or expense is for the account of the goods-owner.

Lord Simon of Glaisdale, who delivered a concurring judgment, took the same view. He thought (p 965E) that to confine agency of necessity to cases where the issue was the bailee’s authority to bind the bailor to contracts with third parties was “justified by the fact that the law of bailment will often resolve any issue between alleged principal and agent of necessity, as it has done here.” The Winson was a decision about the law of bailment. It was not a decision about agency of necessity.

The circumstances which entitle the owners to recover in the present case correspond to those which were decisive in The Winson. They are (i) that the cargo was originally bailed to the owners under a contract which came to an end while the cargo was still in their possession, (ii) that as a matter of law their obligation to look after the cargo continued notwithstanding the termination of the charterparty, and (iii) that the only reasonable or practical option open to them once the charterparty had come to an end was to retain the cargo until it could be discharged at the port where the vessel was then located.

The Winson was a claim for expenses incurred by the salvors, although Lord Diplock’s adoption of the decision in Cargo ex Argos suggests that he would have applied the same principle to a claim for remuneration where the claimant stored and handled the goods with his own facilities. In principle, that seems right. The opportunity cost of retaining the vessel in Angra dos Reis while the charterers’ cargo remained on board was a true cost even if it was not an out of pocket expense. However, it is unnecessary to go any further into that question because in this context as in that of clause 13, no point is taken about the difference between expenses and remuneration.

In the result, I agree with the conclusion reached on this point by Andrew Smith J.

It may well be that in the light of recent developments in this area of law, the owners might be entitled to succeed on this basis also, although the measure of recovery would not necessarily be the same. This, however, raises larger issues which would be better decided in a case where they arise, and possibly in a less specialised context than a dispute about carriage by sea.

I would allow the appeal and restore the order of Andrew Smith J.

I agree, for the reasons given by Lord Sumption, that this appeal should be allowed. I wish only to add a brief explanation of why I agree with him and Lord Clarke that the express indemnity provided by Clause 13 applied to the facts of this case.

I do not view the issue as turning upon a choice between competing causes of the requirement to discharge the cargo. The obligation to discharge a cargo loaded under a time charter will normally be proximately caused by the order to load the cargo. The reason why the consequences of the obligation to discharge are not normally covered by an indemnity clause such as clause 13 of the charter in this case is that those consequences form part of the services that the owners has contracted to provide under the charter and for which hire is being paid. Where, however, the charter comes to an end before the cargo has been discharged in circumstances where the consequences are not expressly covered by the charter, those consequences fall naturally within the scope of the indemnity clause. I accept that the application of the indemnity clause in such circumstances appears to be a novelty, but I can see no argument of principle that precludes this.

I agree with the result reached by the majority, but I do so not under clause 13, but on the basis of the principle in The Winson (China Pacific SA v Food Corpn of India [1982] AC 939), with which Lord Sumption deals in paras 18-30. As he notes (para 29), the charterers in the present case have expressly disclaimed any reliance upon the distinction between reimbursement of expenses and remuneration (as to which, see eg The Principles of the Law of Resitution by Graham Virgo 2nd ed (2006), p 290). They have done this on the basis that the two would on the facts here equate (ie it “cost” the owners the market rate to wait in Angra dos Reis). It is unnecessary to consider the correctness of this concession, and I do not do so.

There is much case law on time charter indemnities. They may be express, as in the time charters which were the subject of Larrinaga Steamship Co Ld v The King [1945] AC 246, 253 and Royal Greek Government v Minister of Transport (The Ann Stathatos) (1949) 83 Ll L Rep 228 and in the Shelltime 3 form of charter in issue in the present case. They may also be implied, as in the case of the New York Produce Exchange form of charter, in which the only relevant express obligation is that the owners or master shall be under the orders and directions of the charterers as regards employment, agency and other arrangements. The existence of an implied time charter indemnity in respect of compliance with charterers’ orders and direction has long been recognised: see The Athanasia Comninos [1990] 1 Lloyd’s Rep 277, 290 per Mustill J, and Triad Shipping Co v Stellar Chartering & Brokerage Inc (The Island Archon) [1994] 2 Lloyd’s Rep 227, 234 per Evans LJ.

The scope and application of an indemnity clause depends upon its precise terms read in the context of the contract as a whole. Other terms of the contract may mean that it is necessarily or impliedly limited in its scope. In addition to that, an indemnity clause in the form of clause 13 will not cover matters of navigation or in respect of which owners can by the contract be taken to have assumed the risk. Within its scope, the present clause also only applies to “consequences or liabilities that may arise from [here] the master complying with charterers’ or their agents’ orders”. This raises a question of causation. The search is for “the ‘proximate’ or ‘determining’ cause”. This was stated in relation to a materially identical clause in Larrinaga Steamship Co Ltd v The King [1945] AC 246, 253 by Viscount Simon LC, with whose speech Lord Thankerton and Wright agreed at pp 253-254.

The issue of causation was considered in depth by Devlin J in The Ann Stathatos 83 Ll L Rep 228. The decision is at the root of the modern jurisprudence on time charter indemnity clauses, and Lord Sumption cites it in para 9. One particular passage is worth citing in full, because it bears on an argument advanced by owners in the present appeal, which the majority might otherwise be thought to be accepting.

In The Ann Stathatos the vessel had been damaged by an explosion resulting from an explosive atmosphere created by the cargo of coal and some unidentified act during repair work causing a flame or spark leading to a series of explosions. The arbitrator selected as “the direct or immediate or effective cause” of the explosions the latter act. Owners argued that it was enough that the explosive atmosphere generated by the cargo was “a” cause. The argument mirrors a submission made by owners on the present appeal, which was rejected in the following passage:

“This conclusion clears the ground for consideration of a further submission on behalf of the owner. The loading, if not the proximate cause, was at any rate, it is argued, a cause of the explosion, and that is sufficient for the purpose of clause 9. Sir Robert Aske does not in this contention rely on the phrase ‘all consequences’; in this I think he is right, having regard to the dictum of Willes J in Ionides v Universal Marine Insurance Co (1863) 14 CB (NS) 259, 289. He relies on the principle applicable in cases of tort, and he referred again to Burrows v March Gas and Coke Co LR 7 Ex 96, though Baron Pigott, in the court below (LR 5 Ex 67, 73) hardly supports the contention. He referred also to Minister of Pensions v Chennell [1947] KB 250, where Denning J discusses the whole matter. As against this, Sir William McNair argues that the term ‘a cause’ can properly be used only when there are two or more causes equal in proximity, as in Reischer v Borwick [1894] 2 QB 548. I need not consider this last contention, for I think it is clear that clause 9 is concerned with the proximate cause. It is a contract of indemnity, and I can see no reason for treating it differently from any other contract of insurance. The observations of Lord Shaw in The Ikaria [1918] AC 350, 368 and the dicta he there cites are also in point.”

The search is therefore for the proximate cause. Devlin J cited Reischer v Borwick [1894] 2 QB 548 as indicating that there can be situations in which two causes are so closely matched that both are identified as proximate causes. That is a largely theoretical analysis which finds little practical application in the authorities, and has achieved any prominence only in discussion about exception clauses. Reischer v Borwick itself was a case on a marine insurance policy covering “only …. collision”, and so not perils of the seas. The vessel was holed by collision, the hole was temporarily plugged, but the plug failed as she was being towed to safety and she sank due to the inflow of water. Not surprisingly, the claim succeeded. Only Lindley LJ addressed the possibility that this situation could and should be analysed as one of concurrent proximate causes (although even he in his concluding remarks identified the injury by collision as “really … the cause of the loss – the causa causans and not merely the causa sine que non”). Both Lopes and Davey LJJ analysed the position throughout in what one would have thought to be more conventional terms as involving a single proximate cause of the sinking (the collision holing the vessel).

Another of the few cases in which courts have discussed the possibility of concurrent causes is Wayne Tank and Pump Co Ltd v Employers Liability Assurance Corpn Ltd [1974] QB 57. The case involved an insurance claim following on from the decision in Harbutt’s “Plasticine” Ltd v Wayne Tank and Pump Co Ltd [1970] 1 QB 447. Harbutt’s factory was burnt down in a fire. The fire occurred because Wayne Tank had installed a pipeline made of unsuitable and dangerous plastic material and wrapped in heating tape attached to a useless thermostat, and had then switched on the heating and left it unattended overnight without testing. Wayne Tank’s policy contained an exclusion of damage caused by the nature or condition of any goods which they sold or supplied. Again not surprisingly, both Lord Denning MR (pp 66G-67B and 68A) and Roskill LJ (p 74 B-C) preferred to analyse the situation as one of a single effective, dominant and proximate clause (the defective plastic material and thermostat supplied), while only Cairns LJ (p 69A) preferred an analysis of “approximately equal” causes. All three member of the Court also indicated that the claim anyway failed (because of the exclusion) even if analysed as one of two concurrent proximate causes (pp 67B-68A, 69B-D and 74D-75E). In both Reischer v Borwick and Wayne Tank, the courts further noted that merely because one can identify concurrent causes does not mean that both are in law proximate causes.

The same point had been made by Lord Shaw fifty years earlier in another leading authority on proximate cause, Leyland Shipping Co Ltd v Norwich Union Fire Insurance Society [1918] AC 350, 370, when he said: “Where various factors or causes are concurrent, and one has to be selected, the matter is determined as one of fact, and the choice falls upon the one to which may be variously ascribed the qualities of reality, predominance, efficiency”. That reasoning was followed and applied in Yorkshire Dale Steamship Co Ltd v Minister of War Transport [1942] QC 691. The issue in that case was whether a vessel lost by stranding in the course of a warlike operation was lost by reason of the warlike operation. Viscount Simon LC said: “Most results are brought about by a combination of causes, and a search for “the cause” involves a selection of the governing explanation in each case” (p 698), Lord Macmillan said: “it is not enough that the casualty arose in the course of a warlike operation. It must also arise out of, and be proximately caused by the warlike operation” (p 702), and Lord Wright underlined the point in a well-known passage, including the statements that “This choice of the real or efficient cause from out of the whole complex of the facts must be made by applying commonsense standards. ….. The question always is what is the cause, not merely what is a cause” (p 706).

Another case involving an exceptions clause where the possibility of rival causes was considered briefly and obiter was Handelsbanken v Dandridge [2002] EWCA Civ 577, [2002] CLC 1227, where in para 47 Potter LJ remarked that “the first task of the court is to look to see whether one of the causes is plainly the proximate cause of the loss” and that “It is only if the court is driven to the conclusion that there was ‘not one dominant cause, but two causes which were equal or nearly equal in their efficiency in bringing about the damage’ one being a period, the other an exception, that the exception prevails”, citing in support Wayne Tank, p 67. That dictum may go further to blur lines than I would in referring to causes “nearly equal in their efficiency”, but, once again, the Court’s actual view was that this was not the situation on the facts. The position regarding exclusion clauses in situations where two causes might be said to be operating concurrently was most recently discussed in Global Process Systems Inc and another (Respondents) v Syarikat Takaful Malaysia Berhad [2011] UKSC 5, para 88. As Devlin J pointed out in The Ann Stathatos at p 237, bottom left, the existence of an exceptions clause is itself likely to affect what falls to be regarded as dominant, proximate or relevant; this is because “the whole of what one might call the area naturally appurtenant to the excepted area must be granted to it”. Indemnity clauses are not subject to such considerations. They cover consequences proximately caused, no more and no less.

This is underlined by another way in which the scope of time charter indemnities is delimited in the case law. Implied time charter indemnities and indemnities like clause 13 apply only where there is “a direct causal link” between the orders and the consequences. The phrase and the emphasis are Lord Hobhouse of Woodborough’s in the leading speech, with which all other members of the House agreed, in Whistler International Ltd v Kawasaki Kisen Kaisha Ltd (The Hill Harmony) [2001] 1 AC 638, 656. Lord Hobhouse made the comment in the course of discussion of the decision in Larrinaga Steamship Co Ltd v The King [1945] AC 246, a case like the present of an express indemnity. He cited in support The White Rose [1969] 1 WLR 1098, another case of an express indemnity. As to implied indemnities: see The Hill Harmony itself and Triad Shipping Co v Stellar Chartering & Brokerage Ltd. (The “Island Archon”) [1994] 2 Lloyd’s Rep 227, 238, where Sir Donald Nicholls V-C noted that “the underlying principle” is that the implied indemnity “extends only to certain consequences flowing from a shipowner complying with charterer’s orders”, one limitation being that “to be within the implied indemnity the loss must arise directly from the charterer’s instruction” (another being that it must also be one which, on a fair reading of the charter-party, the shipowner cannot be taken to have accepted: see para 37 above).

In The White Rose, Donaldson J had the benefit of the formidable advocacy of Mr Anthony Evans for owners and Mr Robert Goff QC and Mr Davenport for charterers. He recited Mr Goff’s submission that one vital element had been omitted from Mr Evans’s case:

“namely, that the right to indemnity only arises if and in so far as the loss suffered by the shipowners can be proved to have been caused by compliance with the time charterers’ instructions” (p 1107).

Donaldson J went on to note Mr Goff’s further observation that causation is rarely a live issue in cases where an owner has on charterers’ instructions signed bills of lading committing him to liabilities over and above his charterparty liabilities, but that causation is all important in other cases.

Donaldson J accepted Mr Goff’s submission, holding that it was “necessary in every case to establish an unbroken chain of causation”, and that:

“A loss may well arise in the course of compliance with the time charterers’ orders, but this fact does not, without more, establish that it was caused by and is in law a consequence of such compliance and, in the absence of proof of such causation, there is no right to indemnity”. (p 1108)

The facts in The White Rose were that a Finnish vessel had been ordered to load in Duluth, Minnesota, where Mr de Chambeau, an employee of charterers’ stevedores was injured while on board. He had left his proper place for purposes unconnected with his work, but owners were liable to him under Minnesota law on the ground that the part of the ship where he had gone lacked fencing. The owners were, it seems, in breach of Finnish law in this respect, but that was expressly disregarded as being irrelevant. Donaldson J nonetheless agreed with the umpire that owners’ indemnity claim failed because “what connected the accident with, and gave rise to, a potential liability and an actual loss was the provisions of Minnesota law”. There was lacking “the necessary causal connection between the order to load and the loss” (p 1108).

The selection of the proximate, determining or, in the more modern terminology, real or efficient cause for the purposes of an indemnity has traditionally been described as involving a “choice … to be made by applying common sense standards as the man in the street or a business or seafaring man would apply them”: The Ann Stathatos 83 Ll L Rep 228, 236 per Devlin J, citing Lord Wright in Yorkshire Dale Steamship Co Ltd v Minister of War Transport [1942] AC 691, 706. Lord Wright’s words were more recently cited under the implied indemnity which was in issue under a voyage charter in Total Transport Corpn v Arcadia Petroleum Ltd (The Eurus) [1998] 1 Lloyd’s Rep 351, 361-362. Such an approach does not, or should not, “conceal, or perhaps reveal” – in Lord Hoffmann’s extra-judicial words giving the Chancery Bar Association lecture in 1999 on “Common Sense and Causing Loss” – “a complete absence of any form of reasoning”. Rather, it should involve a conclusion reached after identifying the relevant context and purpose of the question and the relevant considerations. I do not however regard it as wholly irrelevant that three experienced commercial judges have concluded, without it seems real doubt, that the present indemnity clause does not cover the present case.

Perhaps more striking, since the present constitution is also heavy in commercial experience, is the fact that no previous claim like the present can be identified under any express or implied time charter indemnity; this, despite the fact that time charter clauses entitling owners to withdraw in default of payment of any hire instalment, without anti-technicality provisions, have been commonplace and have given rise to other contentious issues over many past decades. Robert Goff J made no mention of any such possibility in Tropwood AG of Zug v Jade Enterprises Ltd (The Tropwind) [1982] 1 Lloyd’s Rep 45, when considering “the nature of a shipowner’s right to recover from charterers remuneration for services rendered after a ship has been withdrawn from the charterers’ service under a time charter, pursuant to an express contractual right of withdrawal” (p 53). Apart from any express request which might be found to have been made (to render such services), he thought that “their liability (if any) to pay remuneration for the services so rendered can only derive from the principles of the law of restitution”.

Of course, if the owners were bound to third parties by bills of lading which charterers had required them to issue, the continuation of the voyage under those separate bill of lading contracts could engage the time charter indemnity, and could (despite Lord Denning MR’s contrary dictum on appeal in The Tropwind [1982] 1 Lloyd’s Rep 232, 237) lead to charterers having to pay owners the market, rather than the charter, rate. Further, if owners were left with no practical option but to carry the cargo to its destination, then they might still have an argument that their time and money were spent “in compliance with the time charterers’ instructions”. No assistance on this latter situation is derived from the New South Wales Supreme Court decision in J Gadsden Pty Ltd v Strider 1 Ltd (The Aes Express) (1990) 20 NSWLR 57, where the owners failed in a claim against bill of lading holders, who, before the vessel’s withdrawal from charter, had pre-paid freight to charterers under what were charterers’ bills.

The present case differs materially from both these situations. Here, if one asks whether the loss suffered by the shipowners was “caused by compliance with the time charterers’ instructions” - Robert Goff QC’s words accepted by Donaldson J in The White Rose [1969] 1 WLR 1098, 1107-1108 – the natural answer, it seems to me, is: certainly not. It was caused because the charter was at an end, the owners were not performing the charterers’ instructions and they were not receiving hire for the time wasted prior to discharge. The “direct” or “unbroken” causal link required by the authorities is lacking. The loss did not even arise “in the course” of compliance with charterers’ orders, to use Donaldson J’s words quoted in paragraph 70 above. It is true, historically, that no cargo would have been on board but for charterers’ instructions. But that is no test of the proximate or the effective cause, as the authorities make clear: see paragraphs 37 to 47 above. It is also unrealistic to scissor up the instructions between loading and carriage to destination, and to attribute the loss to the instructions to load ignoring the failure to carry. When one engages in such a division, one is in fact recognising that subsequent events superseded charterers’ orders and rendered them a matter of history.

The general contractual context in my view also supports a conclusion that the express indemnity clause is inapt to apply to the present situation. Clause 8 of the charterparty gives owners a simple contractual option. It is accepted that the mere late payment of one instalment did not constitute a repudiatory breach (or a breach of a condition in a sense like that used in the Sale of Goods Act 1979) which could entitle the owners to damages for loss of the charter. That loss flowed from the owners’ exercise of their option to withdraw. The phrase in clause 8 “without prejudice to any claim owners may otherwise have on charterers under this charter” does not create a right of action, and looks on its face only to pre-existing claims. So there is no way in which the time spent discharging in Angra dos Reis can be claimed as damages. Yet it is submitted that, because the owners exercised an option to terminate the charter in mid-flow, the charter indemnity provides them as of right not merely with the charter rate (US45,000 per day), but with the market rate (US158,864 per day) in respect of any delay before the vessel is free to move elsewhere to take advantage of the increased market rate. That would be to give them a claim by way of indemnity for loss they cannot claim by way of damages.

It is also unclear where this submission could or would end. In shipping law certainty is of recognised importance and disputes not to be encouraged. The charter required redelivery at the same port as the port of delivery in the Arabian Gulf (clause 3), with hire being paid up to that point. Logically, the consequences of the charterers’ orders to proceed to and load cargo at Angra dos Reis could, on owners’ case, embrace the whole period during which the vessel was returning to the Arabian Gulf, unless she found other paid employment to take her back. The risk of having to return in ballast to her Arabian Gulf delivery port (or anywhere else) could not be described as an “ordinary commercial risk” which the owners were prepared to accept under this time charter (cf the last sentence of paragraph 16 of Lord Sumption’s judgment), since the owners expressly stipulated against it.

The silence of clause 8 regarding the position post-withdrawal also contrasts with clause 18 which expressly provides that, should the vessel be on a ballast or laden voyage at the date the charter should otherwise terminate, “charterers shall continue to have the use of the vessel” at the charter rate or the market rate if higher. Under clause 18, the charterers are paying for completion of the services requested. Under clause 8, owners, having elected to determine the charter, are now seeking by way of “indemnity” to recover the market rate, without of course having to give any credit for the considerable benefit likely to have accrued to them from such termination.

In conclusion, the majority’s present decision stretches the application of the express charter indemnity beyond any previous decision, without justification, without regard to the potential consequences (including the uncertainty - or certainty - of ever more ambitious claims) and without need. The law is capable of dealing with this situation in a more conventional manner. It will impose on charterers an obligation as bailors to reimburse the owners as bailees for their time and expense spent in looking after the cargo prior to its discharge. It would, even apart from that, probably also impose on charterers an obligation in restitution in respect of any benefit they could be said to have had through the storage on board the vessel of the cargo. But those remedies flow either from the service rendered in that respect by the owners under the compulsion of their legal obligations as bailees, or from the benefit received thereby by the charterers, and not from the express indemnity.

It follows that I too would allow the appeal and restore the order of Andrew Smith J, although I would do so for the reasons and on the basis that he gave, and not those adopted by the majority.

I agree with Lord Sumption that, for the reasons he gives, this appeal should be allowed on the Winson point (China Pacific SA v Food Corpn of India [1982] AC 939). I wish to add a few words of my own on the construction of clause 13 of the charterparty in the light of the sharp difference of opinion between Lord Mance and Lord Sumption.

I have not found this an easy question. Lord Mance makes a powerful case for a narrower application of clause 13 than that preferred by Lord Sumption. His analysis owes much to the approach adopted in a number of decided cases. However, none of them is on facts such as these. As I see it, the question whether the owners are entitled to succeed under the indemnity provided for in clause 13 involves two sub-questions. The first is one of construction of the clause and the second is whether the owners have shown that they are entitled to succeed under the clause on the particular facts of this case, which is essentially a question of causation.

In order to succeed, the owners must show that the expenses (or loss) they sustained as a result of discharging the cargo at Angra dos Reis in Brazil, which was of course the loading port, in the circumstances described by Lord Sumption, were a consequence of their complying with the charterers’ order to load the cargo. The relevant part of clause 13 is in these terms:

“charterers hereby indemnify owners against all consequences … that may arise from the master … complying with charterers’ … orders …”.

I agree with Lord Sumption in paras 10 to 12 that the clause is very wide but that it is neither complete nor unlimited. In particular, I agree with him that the indemnity is not intended to include consequences which are incidental to the service for which the vessel was required to be available under the charterparty. So, for example, it would not include any cost of or in relation to the discharge of the cargo in the ordinary course of events, which would be covered by clauses 5 and 6 of the charterparty, which provide for the services to be provided and paid for by the owners and charterers respectively, or by clause 7, which provides for hire to be paid by the charterers to the owners. Nor, as Lord Mance observes at para 37, would clause 13 cover matters of navigation or in respect of which owners can be taken to have assumed the risk by contract.

I further agree with Lord Sumption that the real question under clause 13 is whether the charterers’ order to load the cargo was an effective cause of the owners having had to bear a risk or cost of a kind which they had not contractually agreed to bear and that, if the charterers’ order was an effective cause in the sense that it was not a mere “but for” cause which did no more than provide the occasion for some other factor unrelated to the charterers’ order to operate, it does not matter whether it was the only effective cause.

It is not I think helpful to use other adjectives to describe the cause. Different adjectives have been used over the years, including “proximate cause”, “dominant cause” and “direct cause”. To my mind they are somewhat misleading because they tend to suggest that the cause must be the most proximate in time or that the search is for the sole cause. Lord Mance says at para 37 that the search is for “the ‘proximate’ or ‘determining’ cause”. However, I respectfully disagree because such a formulation suggests that there can be only one such cause, whereas there may, depending upon the circumstances, be more than one effective cause.

It is true that the cases make some reference to “the determining” or “the proximate cause”. For example, in Larrinaga Steamship Co Ltd v The King [1945] AC 246, 252 Viscount Simon LC said that the proximate cause of the stranding of a vessel was not “warlike operations”. As he put it at p 253, the vessel was attempting to make a voyage without cargo and suffered from a marine peril when doing so. The fact that she was ordered to leave port sooner than her acting master thought was wise could not turn her disaster into the consequence of a warlike operation. He concluded that the “proximate” or “determining” cause was a misfortune in navigation, not attributable to any warlike operation at all. The House of Lords was not considering the possibility of two effective causes.

Lord Mance refers (at paras 38 to 40) in some detail to the decision of Devlin J in Royal Greek Government v Minister of Transport (The Ann Stathatos) (1949) 83 Ll L Rep 228. On my reading of the arbitrator’s findings in that case (as described at pp 231-232) he identified four causes of the first explosion (using the word “caused” in the wide sense of the word). Omitting two causes which are irrelevant for present purposes, the arbitrator found that the first explosion was caused by (a) the loading on board of gassy and dusty coal and the battening down of the hatches so as to trap the gasses and leave coal dust suspended in such air as existed in tween deck space and (d) some act on the part of the crew who were repairing the tanks, which act caused a flame or spark.

Devlin J said at p 237 that by “the wide sense of the word”, the arbitrator meant to include all suggested or possible causes, however remote, and whether causes in the legal sense or not. He added:

“From these five [the arbitrator] selects the act which caused the flame or spark and the explosive atmosphere as the direct or immediate or effective causes of the first explosion.”

The first explosion is the only explosion which is relevant for present purposes. It would seem to follow from that conclusion that there were two effective causes and not one. It would also seem to follow that the arbitrator was choosing causes (a) and (d) as the two effective causes. However the arbitrator then held (as stated at p 232) that “the loading of the coal, while one of the causes of the damage to the ship (using the word ‘causes’ in its wide sense), was not the direct or immediate or effective cause of the loss or expenses claimed”.

It is not clear to me how these findings can be reconciled. One possibility is that the arbitrator treated cause (a) as two causes and not one, by treating the loading of the coal as a different cause from the presence of the gas. If that is correct, the arbitrator held that there were two effective causes, namely the presence of the gas and the flame or spark. It is not easy to see how that is consistent with the view later expressed by the judge that the arbitrator seems to have taken “what is immediate in time”, by which he must have meant the flame or spark (p 237). If the arbitrator treated (a) as one cause, namely the loading of cargo in a gaseous state, it is not easy to see how his conclusion that the flame or spark and the explosive atmosphere were the direct or immediate or effective causes of the explosion is consistent with his conclusion that “the loading of the coal, while one of the causes of the damage to the ship (using the word ‘causes’ in its wide sense) was not the direct or immediate or effective cause of the loss or expenses claimed”.

As I read the judgment of Devlin J, he concluded (at pp 237-238) that the arbitrator favoured the cause that was immediate in time, namely the initial flame or spark which ignited the gas. He rejected the submission that the arbitrator misdirected himself by confusing immediate cause with direct or effective cause. He also rejected (at p 238) the submission that the flame or spark was too remote in law to be the cause of the first explosion. He then considered whether the loading, which the arbitrator rejected as the direct cause, was too remote in law to be a cause at all.

It was in the light of those conclusions that, in the passage quoted by Lord Mance at para 39, Devlin J considered, at p 238, whether, if it was not the proximate cause, loading was a cause of the explosion. This part of Devlin J’s judgment must be set in the context of the facts. Perhaps naturally in the light of the arbitrator’s award, he started with “the proximate cause”, which the arbitrator had held was the flame or spark which ignited the methane gas. He considered the possibility of there being more than one proximate cause, but said that it was not necessary to consider it because the indemnity clause was concerned with “the proximate cause”. He then expressed his conclusion thus:

“So the matter comes down to this, that the arbitrator has selected one cause in preference to another as the proximate or direct cause. I cannot see that any question of law is involved in this selection”.

In short, Devlin J held that that conclusion was a conclusion of fact and that the arbitrator had not misdirected himself in law.

In these circumstances, I do not think that the decision or reasoning in The Ann Stathatos is of any real assistance. The arbitrator had expressly held that there was a sole proximate cause. It may be that, in the light of his earlier conclusion that the direct or immediate or effective causes of the collision were both (a) the act which caused the spark or flame and (b) the explosive atmosphere, the arbitrator made an error in concluding that the spark or flame was the proximate cause, in the sense of sole proximate cause. However, if he did, on Devlin J’s approach it was an error of fact, not an error of law.

In all the circumstances the decision of Devlin J is an unconvincing basis for a conclusion that the search is for the proximate cause. As I see it, the question in each case, whether under a contract of insurance or under a contract of indemnity, is whether an effective cause of the alleged loss or expense was a peril insured against or an indemnifying event. By reference to Devlin J’s citation of Reischer v Borwick [1894] 2 QB 548, Lord Mance accepts in para 64 that two causes may be so closely matched that both are identified as effective causes. However he says that that it is a largely theoretical analysis which finds little practical application in the authorities.

It is true that the authorities do not contain much discussion of the circumstances in which there may be two effective causes. However, in my opinion, they clearly show that two effective causes can, in principle, exist. To my mind this can be clearly seen from Wayne Tank and Pump Co Ltd v Employers Liability Assurance Corpn Ltd [1974] QB 57, Lloyd (JJ) Instruments Ltd v Northern Star Insurance Co Ltd (The Miss Jay Jay) [1987] 1 Lloyd’s Rep 32 and Midland Mainline Ltd v Eagle Star Insurance Co Ltd [2004] EWCA Civ 1042, [2004] 2 Lloyd’s Rep 604.

The present position can be most clearly seen from the Midland Mainline case, where Sir Martin Nourse, with whom Brooke and Jacob LJJ agreed, expressly held at para 48 that there can be more than one proximate cause of loss. He cited Leyland Shipping Co Ltd v Norwich Union Fire Insurance Society Ltd [1918] AC 350, Wayne Tank and The Miss Jay Jay as authority for that proposition.

It is true that in Wayne Tank [1974] QB 57, on the facts (which are described by Lord Mance at para 41) the majority of the Court of Appeal, Lord Denning MR and Roskill LJ, held that the proximate cause of the fire was the defective plastic material and thermostat supplied and not the act of switching on the heating and leaving it unattended without testing. However, in a passage quoted by Sir Martin Nourse at para 10 of the Midland Mainline case, Roskill LJ said that he found it impossible to say that the latter was the sole proximate cause of the fire and, that if he was wrong to say that the defective state of the material and thermostat was the sole proximate cause of the fire, there were two effective proximate causes. Cairns LJ, whose approach Sir Martin described as different but instructive, said at p 68:

“But for my part I do not consider that the court should strain to find a dominant cause if, as here, there are two causes both of which can be properly described as effective causes of the loss. Mr Le Quesne recognised that if there are two causes which are approximately equal in effectiveness, then it is impossible to call one rather than the other the dominant cause. I should prefer to say that unless one cause is clearly more decisive than the other, it should be accepted that there are two causes of the loss and no attempt should be made to give one of them the quality of dominance.”

Those were cases in which it was held that, where the or a proximate, or effective, cause of the loss is excepted by the policy, the insurers are not liable. It is, however, clear from The Miss Jay Jay that, where there are two effective causes, neither of which is excluded but only one of which is insured, the insurers are liable. In the Court of Appeal Slade LJ underlined (at p 39) that the authorities show that the question of proximate cause has to be determined by “a broad commonsense view of the whole position” and that, by proximate, is meant proximate in efficiency. It was held that the faulty design and construction of the vessel, which was neither an insured peril nor an excepted cause, and perils of the seas, which was an insured peril, were both proximate causes of the loss since they were, as Slade LJ put it at p 40 “equal or at least nearly equal in their efficiency in bringing about the damage”. These principles are as I see it correctly summarised in McGillivray on Insurance Law, 11th ed (2008) at para 19-005 under the heading “Two effective causes” and in McGee on The Modern Law of Insurance 3rd ed (2011) at pp 260-261. See also to the same effect McCann’s Executors v Great Lakes Reinsurance (UK) Plc [2010] CSOH 59, para 112 to 117, where Lord Hodge also stressed the importance of context; Orient-Express Hotels Ltd v Assicurazioni General SpA (UK branch) (trading as Generali Global Risk) [2010] EWHC 1186 (Comm), [2011] Bus LR 7 per Hamblen J; and Global Process Systems Inc v Syarikat Takaful Malaysia Bhd [2011] UKSC 5, [2011] Bus LR 537, para 88 per Lord Mance and, in the Court of Appeal, [2009] EWCA Civ 1398, [2010] 2 All ER 248, para 32 per Waller LJ.

I entirely agree with Lord Mance that there must be a causal link between the order and the consequences relied upon. In short, there must be no break in the chain of causation between the order and the consequences. This is clear from The White Rose [1969] 1 WLR 1098. As Lord Mance says at para 45, Donaldson J there accepted that it was “necessary in every case to establish an unbroken chain of causation”. That is the sense in which I read Lord Hobhouse of Woodborough’s reference to the necessity for a “direct causal link” in Whistler International Ltd v Kawasaki Kisen Kaisha Ltd (The Hill Harmony) [2001] 1 AC 638, 656. Lord Hobhouse was not considering a case like the present. I do not read him as intending a direct causal link to be different from an effective cause. I remain of the view expressed above (and in agreement with Lord Sumption) that the question is whether the relevant order was an effective cause of the alleged consequence.

I agree with both Lord Sumption and Lord Mance that in deciding whether causation was established on the facts, it is important to have in mind the context in which the question is asked. I do not think that the answer can be found in the conclusions on the facts to which Lord Mance has referred. In particular, I do not think that Donaldson J’s conclusion based on the vagaries of Minnesota law in The White Rose is of any assistance in deciding the relevant question of fact in this appeal.

Lord Sumption has identified the relevant commercial context. It is that the charterers gave an order to load the cargo in the ordinary way. The consequence of that order was that the cargo was indeed loaded and therefore had sooner or later to be discharged. The failure of the charterers to pay hire timeously triggered the owners’ right of withdrawal. When they exercised that right, the question was what should be done with the cargo which was still on board the vessel. It had to be discharged somewhere. As it happened, it was discharged at the port of loading but it might have been discharged at the port of discharge or at an intermediate port. I entirely agree with Lord Sumption’s analysis at paras 9 to 16. In particular I agree with Lord Sumption that the owners’ motive for exercising the right to terminate is irrelevant.

It was adventitious where and when the termination occurred. The position would have been the same if the termination had occurred for some other reason than the exercise of an option by the owners, as for example as a result of frustration. The owners would have had to procure discharge of the cargo and would have incurred expenses and perhaps loss. They would not have been able to recover such expenses and loss under any of the other provisions of the charterparty. The reason they would have to incur the expenses is that the cargo was still on board the vessel. Just as here there are two effective causes of the expenses and loss, namely the withdrawal and the fact that cargo had been loaded, so in a frustration case, there would be two such causes, namely the frustration and the fact that cargo had been loaded and was on board.

I agree with the view expressed by Lord Mance at para 50 that, if the owners were bound to third parties by bills of lading which charterers had required them to issue, the continuation of the voyage under those bill of lading contracts could engage the indemnity under clause 13. Lord Mance further recognises (to my mind correctly) that if owners were left with no practical option but to carry the cargo to its destination, then they might have an argument that their time and money were spent “in compliance with charterers’ orders”. Indeed, at present I see no reason why they should not succeed under the indemnity in such circumstances. As I see it, that would be on the basis that the charterers’ orders would be the orders to load. In terms of causation, I see no distinction in principle between that case and the present.

For these reasons and the reasons given by Lord Sumption I would allow this appeal on the indemnity point as well as the Winson point.