IF P & C Insurance Ltd (Publ) v Silversea Cruises Ltd & Ors

On 11 September 2001 (“9/11”), as is well known, four American civil airliners were taken over by force by Islamic fundamentalist adherents of the Al Qaeda organisation. Two of the planes were piloted into the twin towers of the World Trade Centre in New York and a third into the Pentagon in Washington DC, while the fourth, which it is thought may have been intended for the Capitol or the White House, crashed in the Pennsylvanian countryside. The impact of 9/11 was worldwide. Among the commercial fall-out was a reluctance by customers of luxury cruises to travel, either to particular destinations or at all. The proceedings out of which this appeal arises concern an insurance claim made as a consequence of 9/11 by the operators of a fleet of four ultra-luxury cruise ships under a policy designed to protect them against loss of income.

The policy is dated 9 April 2001 and was effective for 12 months from 1 April 2001. It is called a “Loss of Income and Extraordinary Costs” policy. It was made between IF P&C Insurance Limited, who are claimants and in this appeal respondents (the “underwriters”), and the assured, in these proceedings the defendants and here the appellants, who are described in the policy as –

As of April 2001 there were three vessels within the Silversea fleet, although a fourth was anticipated to come into operation within the near future. The three vessels were The Silver Cloud, The Silver Wind, and The Silver Shadow. The Cloud and the Wind were older and smaller than the Shadow, which had joined the fleet in September 2000 as a new vessel. The fourth was The Silver Whisper, another new vessel, and a sister-ship to the Shadow. The Whisper started service on 24 June 2001 and as a result the original 9 April 2001 policy was amended and re-issued in June 2001, still dated 9 April 2001, so as to incorporate reference to her “Effective June 24, 2001 until common expiry”.

Thus the parties to the policy as assured, and hence to these proceedings and this appeal, include Silversea Cruises Limited, the holding and management company which is a Bahamian company with headquarters in Fort Lauderdale, Florida, and the four shipowning companies themselves. Unless it is necessary to distinguish between the assured companies, like Tomlinson J, who was the trial judge below, I shall refer to them simply as “Silversea”.

The full terms of the policy are annexed to this judgment at Annex 1. For the present, in order to give focus to an account of the relevant facts and issues, it is sufficient but important to concentrate on the critical wording concerning the insured perils. There were three separate covers, Ai, Aii and B, albeit Ai and B are closely connected.

The heading of Ai referred to “Loss of income” and the cover was described as –

These provisions concerning the nature of Ai cover were in fact introduced as an amendment to the Norwegian Plan, to which it is necessary to refer.

The Norwegian Plan is a compendium of standard provisions for the insurance of sea-going oil tankers. The Silversea policy incorporated an amended form of chapter 16 of the 1996 version of that Plan, the full title of which is the Norwegian Marine Insurance Plan of 1996. The incorporation applied to cover Ai and was effected by the following words –

The words of incorporation make it clear that §16-1 of chapter 16 was replaced by the policy wording, but that the remaining provisions of chapter 16 are to be incorporated. Chapter 16 is headed “Loss of hire insurance”. §16-1 sets out the scope of the cover for oil tankers (“loss due to the vessel being wholly or partially deprived of income as a consequence of damage to the vessel…also…because it has stranded…because it is prevented by physical obstructions from leaving a harbour…[or] as a consequence of measures taken to salvage or remove damaged cargo”), and the remaining provisions of chapter 16 describe how any indemnity in question is to be calculated. For present purposes the essence of those provisions emphasises that underwriters’ liability is to be calculated by reference to “loss of time” (§16-3), that such loss of time is to be calculated in days, hours and minutes and limited to the “sum insured per day multiplied by the number of days of indemnity per casualty and in all” (§16-4), that if a fixed amount per day is written into the policy, then “this amount shall be regarded as an assessed insurable daily amount unless the circumstances clearly indicate otherwise” (§16-6), and that each casualty shall be subject to a deductible period to be reckoned from the beginning of the casualty “until the loss of time… reaches the number of deductible days stated in the policy” (§16-7).

The intention of the Plan appears to be that loss of hire (the nature of an oil tanker’s income) is to be measured by loss of time and that where there is damage to the vessel or some other peril which causes loss of time greater than the deductible, then the vessel owner’s indemnity is assessed as a straightforward calculation relating to an agreed amount per day for all further loss of time up to the policy limit. It will be seen (infra) that the Silversea policy similarly provides, in relation to cover Ai, for a daily rate subject to a deductible and a limit.

The heading of Aii referred to “Loss of anticipated income…” and the loss was further described as

That loss was still further described (I again use underlining to emphasise the critical words relied upon) as –

Thirdly, the heading of cover B referred to “Cruise Credits/On-Board Credits” and the loss was further described as –

That loss also underwent further description, in this case –

Although cover B comes after cover Aii, its close relationship with cover Ai is emphasised not only by its wording “Subject to the same conditions as in A.i.) above…”, but also by the fact that the premium for covers Ai and B is rated together, whereas the premium for cover Aii is separately provided for.

The premium for covers Ai and B is described under the heading “Rate” and set as “1.43% per annum and pro rata”. Although the policy did not expressly say so, that percentage rate of 1.43% attaches to the maximum sum insured in respect of each of the four vessels. Those sums are contained in a “Sum Insured Schedule (applicable to Section 1.A)” (sic, sc Section Ai) and varied for each vessel to reflect its size and earning capacity, from about $8/$9 million in the case of the Cloud and the Wind up to over $15 million in the case of the Shadow and the Whisper.

The premium for cover Aii, however, was agreed as a lump sum to cover the fleet of four vessels as a whole. The agreement was $175,000 to cover the Cloud, Wind and Shadow with an additional $25,000 (to make the figure up to $200,000) once the Whisper entered the fleet. Thus in the earlier policy which only covered the Cloud, Wind and Shadow, the cover Aii premium is expressed simply as “US$175,000”. In error, at the invoicing stage, the cover Aii premium was billed at $58,333.33 per vessel ($58,333.33 x 3 = $175,000). That error was then carried forward into the later policy which provided that the cover Aii premium was “US$ 58,333.33 per annum and pro rata”, which would of course have given a total premium well in excess of $200,000. Before the trial judge Silversea sought to maintain that the premium provision in the later policy was not an error and that the parties’ agreement had always been for a premium per vessel rather than a single lump sum premium for the fleet (adjusted to take into account the fleet’s expansion from three to four vessels). However, that argument is no longer pursued on this appeal. In short, therefore, the cover Aii premium can be regarded as originally providing for $175,000 and ultimately providing for $200,000.

It is important at an early stage in this judgment to point out that provisions for “Limitations and Deductibles” operated separately in relation to each of the three covers. The relevant terms are set out in Annex 1, but I would emphasise the essential provisions, for to a large extent they have driven the legal disputes between the parties.

Thus in relation to cover Ai, the limitation was expressed as –

The “attached” was a reference to the Sum Insured Schedule to which I have already referred. This contained not only maxima, but also “per diem” amounts which varied from one vessel to the other and also within various cruising seasons. Thus the year was divided into three seasons and, for instance in the case of the Cloud, her per diem figures varied from $75,000 for the low season up to $115,000 for the high season. These per diem figures were clearly intended to reflect the value to the assured of a day’s cruising for each vessel in the appropriate season.

The maximum sums insured set out in the Schedule, to which I have already referred, meant that across the fleet as a whole a total indemnity of over $47 million was possible under section Ai.

In relation to cover Aii, however, there is a limit of $5 million: the clause did not refer to the Schedule but simply said:

A critical issue in this appeal is whether that $5 million is a limit per vessel or across the fleet: on that issue depends whether the policy responds under section Aii in a potential total of only $5 million or up to $20 million.

In relation to cover B, the clause simply provides:

As for Deductibles, there is a combined deductible for covers Ai and B –

In this respect as well, therefore, Ai and B show their inter-relationship. This deductible, being stated in terms of days, fits closely with the Sum Insured Schedule as well as with the logic of chapter 16 of the Norwegian Plan. Cover Aii’s deductible, on the other hand, is not stated in terms of days but money –

Two further provisions need to be highlighted. One is within the “Exclusions” clause and provides:

Silversea rely on this exclusion in support of their argument that cover Ai is apt to provide an indemnity for loss of market causing passenger and cruise cancellations. The underwriters also rely on this exclusion, raising a new argument not addressed at trial, to the effect that it excludes any liability under the Aii cover. It will have been observed that whereas the Ai cover, and thus the B cover too, are triggered by perils such as terrorism, the Aii cover is triggered by “a State Department Advisory or similar warning”. The underwriters therefore submit that, even if such warnings have contributed to a cover Aii loss of market, the resultant losses are excluded if they have also been caused by anything other than such warnings, eg by terrorism itself.

The other provision falls under the “General Terms and Conditions” (applicable to All Sections), viz –

The judge held and ordered that Silversea’s claims under cover Ai and cover B of the policy were dismissed (save in so far as they related to a claim arising out of mechanical breakdown of the Cloud in October 2001, a claim which was separately settled by the parties), but that its claim under section Aii succeeded in the maximum sum of US$5,000,000.

On this appeal Silversea seek to reinstate their claims under sections Ai and B and also to argue, as they did below, that the limit under section Aii is not $5,000,000 across the fleet as a whole, but is $5,000,000 per vessel. That is probably the most significant single issue, in terms of the parties’ disputes, in this litigation. If the limit under section Aii were indeed $5,000,000 per vessel, then it may well be that Silversea would be less concerned whether they could bring their claimed losses also within sections Ai and B. The underwriters, on the other hand, have accepted for the purposes of this appeal that they are liable for $5,000,000 under section Aii. That concession, however, is subject to this reservation: that upon the true construction of the exclusion mentioned above, there has been no insured loss at all. The reservation is said to be made to protect them against the possibility that Silversea are right to say that the Aii limit of cover is $5,000,000 per vessel. The underwriters state that even if they are right about the effect of the exclusion, as well as being right about the limit of cover, they are nevertheless prepared to maintain their acceptance of Aii liability in the amount of $5 million.

Aii cover: I have already said that probably the most significant issue in this court is (i) whether the limit under the Aii cover is $5 million per vessel or across the fleet as a whole. If this appeal were to confirm the judge’s decision that the limit is $5 million across the whole fleet, then, in the light of the underwriters’ undertaking to accept Aii liability in the sum of that limit, it is common ground that no other issue that arises in connection with the Aii cover is of any financial materiality. However, it is convenient at this stage to expose and list the further issues which have been argued. They are (ii) whether the policy should be rectified to clarify that the $5 million limit is a fleet and not a vessel limit; and (iii) whether the six months time limit contained in cover Aii (“subject to a maximum period per event of 6 months”) is a period for identifying the cruises in respect of which a fall-off in passenger bookings is to be measured or alternatively a period in which the fall-off in bookings is measured for all cruises advertised in the “Cruise Atlas” current at the date of the policy. The further issue referred to above in relation to the exclusion clause, namely (iv) whether market losses due to 9/11 itself are excluded, even if also due to government warnings, is only material if the $5 million limit is a per vessel limit.

Ai cover: The other significant issue relates to whether Silversea have any claim at all under cover Ai. This is a complex issue which brings in its train a number of sub-issues, but the essence of it is (v) whether loss of passenger bookings, whether or not accompanied by voyage cancellations, is within cover Ai at all. Silversea say that it is, being “loss due to the vessel being wholly or partially deprived of income as a consequence of…terrorists” within the fifth peril and/or “…acts of war, armed conflict” within the seventh peril. The underwriters say it is not since, inter alia, the policy structure provides for loss of market due to loss of passenger bookings to be compensated under cover Aii not cover Ai at all, and that cover Ai is rather concerned with physical interference with the vessels rather than with loss of bookings. This argument brings with it the following sub-issues: (vi) whether Silversea’s Ai claim, expressed in terms of loss of passenger bookings, can survive in the face of the Ai cover requirement of an interference with a scheduled itinerary (both the fifth and the seventh peril refer to events which interfere with “the scheduled itinerary of the insured vessel”); (vii) whether Silversea’s Ai claim, expressed in terms of loss of passenger bookings, can survive in the face of the Norwegian Plan’s requirement of a loss of time; (viii) whether Silversea has preserved an alternative formulation of its Ai claim in terms of the per diem loss of time effect of voyage cancellations due to the lay up of the Wind; and (ix) whether in this connection it can be said that 9/11 was not merely “terrorism” within the fifth peril (that is common ground) but also amounted to “acts of war” or “armed conflict” within the seventh peril.

B cover: The claim under cover B raises no separate issues, for it is parasitic on cover having been established under Ai and the joint deductible having been surpassed.

In essence therefore there are two groups of issues triggered by Silversea’s claims. As for cover Aii, which provides an indemnity for loss of market, the underwriters concede liability for $5 million, but no more, whereas Silversea seek an indemnity limited only by figures of $5 million per vessel. As for cover Ai, the underwriters deny any liability at all: they say that Silversea are seeking a market loss under a cover which only responds to physical interference with consequent loss of time; whereas Silversea are seeking an indemnity for passenger and cruise cancellations. Presumably Silversea do so, in part from a concern that they are too narrowly limited to only $5 million under cover Aii. On both these essential matters, the judge has decided in the underwriters’ favour.

It is unnecessary at this stage to set out the events of 9/11 further. The parties were content to accept that the events, the perpetrators and their motivation, the context and consequences were all accurately described in the reports prepared on behalf of Silversea by Professor Paul Wilkinson and Dr Rohan Gunaratna jointly, both of the Centre for Study of Terrorism and Political Violence of the University of St Andrews.

In the aftermath of the atrocity the US authorities issued a series of warnings. They are relevant because, as the wording of section Aii has shown, such warnings were an insured event. The first of them, described as a Worldwide Caution, was issued immediately on 12 September 2001 by the US Department of State and included the following language:

The judge also cited another Worldwide Caution issued on 23 October 2001, which included the following:

As appears from the second notice cited above, the invasion of Afghanistan by the US and other Western nations commenced on 7 October 2001 as a direct result of 9/11. There were many other such warnings and governmental pronouncements, the constant theme of which was the likelihood of additional terrorist activity directed towards American citizens and interests, as well as of the likelihood of retaliation against American targets in the event of the foreshadowed and actual invasion of Afghanistan.

In the meantime, Silversea reacted promptly to the new situation. On 17 September 2001 Silversea’s senior vice president of marketing, Mr Bill Leiber, initiated a strategic reconsideration of the 2001, 2002 and 2003 itineraries of the fleet “in light of a diminishing area for “safe” travel”. He recommended re-deployment away from areas such as the Suez Canal, Arabian Sea and Indian Ocean and into the Americas and Caribbean. He contemplated that “depending on military offensive” one or more ships might have to be taken out of service. The judge referred to Mr Leiber as an impressive witness.

I summarise the findings of the judge as to the impact of 9/11 and the US warnings on Silversea’s vessels and market. The context of these findings was a trial of liability only. Thus the considerable detail relevant only to matters of quantum was left over as might be necessary for another occasion. The immediate impact on Silversea was that the port of New York was closed on 11 September and remained closed for over six months. This necessitated the substitution of Philadelphia for New York on a cruise by Whisper which had left Dublin on 11 September itself (voyage 4107), and on her three subsequent cruises (voyages 4108/4109/4110). The latter three voyages gave rise to a claim of loss arising out of passenger cancellations ($276,330) and the granting of cruise credits ($744,622, but this in relation to voyage 4108 only). The judge described these claims, the detail of which could await a trial of quantum, as an example of “an enforced change to four scheduled itineraries (Footnote: ¹) in consequence of closure of a port and, indeed, on any view, in consequence of terrorist activities” (at para 10 of the judgment below, where he was presumably responding to the wording of the fifth peril under covers Ai and B).

There were also other minor changes to the itineraries of the vessels on at least some of their voyages in 2001 and 2002: but it has not been submitted that the judge was wrong to conclude (see para 11 of the judgment below) that those minor alterations have not been alleged to have given rise to any loss or any claim. In any event there has been no attempt on this appeal to demonstrate that those minor alterations, even if properly to be described as giving rise to interference with scheduled itineraries, have caused any loss in their own right.

The essential basis on which Silversea have built their claims therefore is the loss of business which 9/11 and the subsequent US government warnings engendered. It is not I think in dispute that there was a loss of business, although the extent of that loss is undoubtedly controversial. However, it is recognised in general, for instance in this passage from the judge’s judgment (at para 9):

Silversea’s reaction to these problems was speedy. On 28 September 2001 Mr Leiber wrote to their agency clients to inform them of their decisions: namely, to lay up one of their four vessels, the Wind, for the rest of the 2001 season and for the whole of the 2002 season and to take that opportunity to refurbish her; to re-deploy the newly refurbished Cloud to perform the Wind’s voyages; to maintain the 2002 itineraries of the new vessels, Shadow and Whisper; and to offer their passenger clients who were already booked but were affected by any of these new arrangements to receive either a full cash refund or a double (“two-for-one”) cruise credit for rebookings.

Silversea also relaxed its published Terms and Conditions in favour of passengers who cancelled voyages due to take place in the immediate aftermath of 9/11. Thus all passengers booked to sail on cruises departing between 11 and 30 September, and some passengers on Wind voyages departing in October, were offered cruise credits in an amount equal to the cancellation penalty otherwise payable by them (but no cash refunds). The judge said that he was satisfied that such steps were prudent in Silversea’s own interests, genuinely believed to be appropriate to the circumstances, and not motivated by any belief that the cost would be borne by the underwriters. These factors may explain why the cruise credits claimed under cover B are as large as they are (see below), but, since any claim under cover B is in general parasitic on cover being established under Ai the size of the claim in itself does not assist Silversea unless the Ai issues are answered in their favour.

As for the cover Ai claim, without going for the moment into how it is made up and quantified, I would observe that it is premised in Silversea’s amended statement of case on two main allegations of fact. The first is that 9/11 created a situation where there was demand for only three out of the four vessels – hence the need to reduce the number of cruises and vessels offered by laying up the Wind. The second is that the Cloud’s scheduled voyages were cancelled as well, in favour of using the Cloud to perform the Wind’s scheduled voyages, because the Cloud’s scheduled voyages were to unsafe destinations in the Middle and Far East. Thus Silversea’s pleadings contained these two critical averments (at para 22.1 of Silversea’s “consolidated document”):

as a result of 9/11, cancellations of bookings for cruises and reduction of business were such that Silversea had insufficient bookings to employ all four vessels and it was therefore necessary to lay up one of the vessels. It was decided by Silversea to lay up one of the older vessels which had not been recently refurbished, the SILVER WIND. That vessel has been laid up since 2^nd November 2001 and is likely to have to remain laid up until the end of May 2003;

cruises which should have taken place on the SILVER WIND during the period of lay-up did not take place on the SILVER WIND, but on the SILVER CLOUD instead, beginning on or about 20^th December 2001. The decision was taken to use the SILVER CLOUD to perform the SILVER WIND’s cruises because the SILVER CLOUD was a more modern vessel and the cruises which had been due to take place on the SILVER CLOUD included destinations in the Middle and Far East (in relation to which substantial numbers of cancellations occurred as a result of 9/11);

accordingly, those cruises which were scheduled to take place on the SILVER CLOUD could not (and did not) take place.”

Now as to those two critical factual allegations, the judge found as follows. First, as to the need to cancel cruises on the ground of passenger reaction to 9/11, there was a warm debate as to whether the true basis for Silversea’s decision to take one vessel out of circulation was not due to the effects of 9/11 but rather to a commercial situation of over-capacity which existed even prior to 9/11. That debate centred around the validity of the way in which Silversea sought to demonstrate its loss of market for the purposes of its cover Aii claim, by reference to the so-called “Bajaj curves” which were an attempt to project future bookings in the absence of 9/11. The judge cast some even-handed aspersions on this subject. He criticised Silversea for their failure to tell the underwriters in advance of a point which only emerged on day 3 of the trial, which was that the curves had been adjusted, and also observed that at any rate the contemporary perception had been that “in terms of bookings Silversea was not doing particularly well” (at para 31). On the other hand, he also expressed his surprise that underwriters had felt able to observe in their closing submissions that the Bajaj curves were “an obviously contrived document deliberately manipulated to present a false picture” (at para 32). Perhaps significantly, at any rate the following has been much relied on by the underwriters on this appeal, he observed (at para 35) that –

Ultimately, however, the judge was spared the burden of making findings about this dispute by the parties’ agreement that it need not be resolved at that trial. The background and motivation of this agreement are somewhat obscure. It appears to have been engendered by the thought that the court’s decisions as to the proper construction of the policy might well make the resolution of such factual matters unnecessary. However, the agreement is plainly referred to by the judge (at para 45):

This agreement is moreover acknowledged in the skeleton argument on behalf of the underwriters for this appeal, where Mr Michael Swainston QC and Mr Alan Maclean refer to the question of over-capacity and say (at paras 27/28):

I mention this matter because Mr Swainston at one point, and by reference to the passage cited above from paragraph 35 of the judge’s judgment, submitted that Silversea’s case on appeal under cover Ai was simply beside the point in the absence of a finding below that the cancellation of the voyages of the Wind had been caused by commercial exigency in the wake of 9/11. In my judgment, however, in the light of the parties’ agreement the absence of findings on what I have described as the first of Silversea’s two main factual allegations in support of their cover Ai claim does not absolve this court of the responsibility of determining the legal points of construction.

The judge did, however, determine the second of the two points, that relating to the safety of the Cloud’s original itinerary. Silversea relied on this to explain the cancellation of Cloud’s voyages so as to free her to perform the voyages which had been scheduled for the Wind. As to that, the judge said this (at para 45):

There has been no challenge on this appeal to that finding. It follows that, if Silversea are ultimately to succeed on their cover Ai claim so far as cancelled voyages are concerned, it would have to be solely on the basis that the commercial decisions to cancel those voyages were taken to deal with over-capacity caused by 9/11 (an allegation yet to be established, but reserved for some future hearing, if necessary), and not on the basis that the scheduled voyages themselves could not take place because of safety fears.

The trial below was not concerned with quantum, and neither is this appeal, but it is not possible to put the arguments of construction which arise into their appropriate context, or to understand some of the issues, without reference to the structure of Silversea’s claims as they have altered over the course of this litigation. The judge himself commented (at para 21):

However, the structure of the policy is not a shining example of clarity, complex claims have had to be advanced before the facts are found by the court, and it is sometimes only in the clash of competing submissions and after room for careful analysis that critical matters are clarified. In the end the claims and their alternatives have to be dealt with on their own merits.

On 9 November 2001 Mr Bajaj wrote a claim letter to Silversea’s New York brokers, The Berkely Group. The judge described it as detailed and considered. It was accompanied by various schedules setting out monetary claims in respect of each vessel under each cover. In due course it was incorporated in Silversea’s defence and counterclaim served on 21 February 2002.

The Ai claim: was essentially in respect of the Wind, by reason of its lay-up from 2 November 2001 to the end of 2002. It was calculated on the appropriate per diem rates set out in the Sum Insured Schedule. The claim amounted to over $44 million but was capped at the Wind’s limit of $8,812,530. There was no 9/11 claim in respect of the Cloud, only in respect of its mechanical breakdown in October 2001. That claim, which as I have mentioned has been settled, was also calculated on a per diem basis over the period from 11 October to 20 December 2001. The only claim in respect of the Shadow was for expenses, claimed by reference to additional wording under cover Ai – “including extraordinary expenditure incurred to prevent or minimize loss of income covered by this policy” – with which I have not had to be concerned since this and all similar expenses claims have since been abandoned. The only claim in respect of the Whisper (apart from a similar claim for expenses, also abandoned) was for $44,764 loss of income due to passenger cancellations on voyage 4109 arising out of the closure of the port of New York. It may be noted that this last item does not seem to fit easily with the structure of the rest of the Ai claim.

The Aii claim: Silversea compared their anticipated bookings as at 9/11 (in fact as at 5 September 2001, the date of their most recent forecast) with actual bookings in respect of voyages which had commenced between 9/11 and the date of the claim letter, and with estimated bookings “over the next six months” for the remainder of the 2001/2002 scheduled voyages. Save for comparatively small “bonus commission” losses, it would seem that the Wind’s losses were dealt with under the Ai claim. In the case of the other three vessels, however, the Aii claim exceeded, and was severally capped at, $5 million.

It is unnecessary to go into the November 2001 cover B claim, save to remark that it was said to reflect cruise credits in respect of passenger cancellations. In the case of the Cloud (some $1.7 million), the Shadow (some $2.9 million) and the Whisper (some $0.75 million, the figure mentioned at para 41 above) the claim was generated by cancellations on voyages due to commence in the remaining part of 2001. In the case of the Wind, however, it embraced not only passenger cancellations (some $1.5 million) in respect of the remaining voyages performed in 2001, but also a further sum of nearly $20 million, capped at $5 million, in respect of the cancelled voyages during her period of projected lay-up. The judge was right to call this a “surprising” claim, since it appears to have been already covered under the Ai claim, and seems to have been uninfluenced by the fact that the Cloud was to perform the Wind’s itinerary in any event.

In essence, however, the points to note about the November 2001 claim are that (i) it was geared towards recouping Silversea’s loss of income from laying-up the Wind, (ii) the Wind’s (and the Cloud’s mechanical damage) income losses under cover Ai were computed by reference to the Sum Insured Schedule on a per diem basis, (iii) passenger cancellations appear to have been claimed under cover Aii and/or cover B but not under cover Ai, (iv) the $5 million limitations in respect of both covers Aii and B were applied on a per vessel rather than a fleet basis, and (v) the six months limitation under cover Aii was applied to a booking period rather than a cruise departure period.

In October 2002 Silversea’s solicitors delivered to the underwriters’ solicitors a new claim letter (to an extent foreshadowed, at any rate with hindsight, in further information served in July 2002). This led to the court requiring the parties to serve afresh “consolidated” statements of case. Silversea’s is contained in their “consolidated document”, to which I have already referred. Tomlinson J rightly described Silversea’s new claim as “radically different” (para 21).

The Ai claim: It is difficult to encapsulate the logical structure of the trial claim and its difference from the November 2001 claim in one overarching sentence, especially if one wishes to avoid all the dangers of over-simplification. Nevertheless it is important to try. My attempt is to observe that the Ai claim was now constructed not on the basis of loss of time calculated on a per diem basis in respect of the Wind voyages which had been cancelled, but on the basis of the net value of passenger cancellations in respect of voyages most of which had been performed. Qualifications and/or clarifications, however, immediately have to be made. (i) The Cloud claim arising from the October 2001 mechanical damage was maintained on a per diem basis ($6.2 million). (ii) The passenger cancellation claims in respect of the Cloud (a further $4.7 million) ranged across all her remaining scheduled voyages, most of which however had been cancelled in order to free her to perform the itinerary of the Wind. (iii) The per diem claim for the Wind’s cancelled voyages was now reformulated as a smaller claim ($2.5m) in respect of passenger cancellations on those cancelled voyages: I assume those are passengers who did not simply switch their reservations to the Cloud which of course performed the Wind’s itinerary. (iv) The Shadow ($5.2 million) and the Whisper ($5.9 million) now generated large claims: yet all their voyages were performed. (v) In respect of the voyages which were performed, there were some minor changes to itineraries, for instance the replacement of turnaround ports or the cancellation or replacement of certain port calls. (vi) Net value of cancellations was arrived at by adding estimated lost on board revenue to the ticket revenue and then deducting saved hotel and food expenses. (vii) The combined Ai and B deductible (10 days) was shown as having been used up in the case of any vessel by the length of the first voyage in respect of which any passenger cancellations arose: thus for instance on the Shadow the 10 day deductible was shown as having been exhausted by reason of the $5,294 claim on the 11 day voyage 3121 commencing on 19 September 2001, and the sum of $5,294 was deducted from the total.

Although this reformulated passenger cancellation claim was no longer put on a per diem basis, Silversea’s consolidated document emphasised that the original per diem claim had not been abandoned. Thus at para 36(2)(a) and (b) Silversea pleaded –

No further details of the per diem rate were there given, but for that very reason it may properly be inferred that the alternative per diem claim is a reference to the calculations contained in and annexed to the 9 November 2001 claim letter. In Silversea’s earlier defence and counterclaim those calculations and schedules were simply incorporated by reference to the claim letter. At any rate the unparticularised form of this pleading, plus in all probability the manner in which the Ai claim was dealt with at trial, led Tomlinson J to consider that the per diem claim had simply been abandoned (para 21). This has itself become an issue on this appeal (issue (viii)), and I shall deal with it further below.

The Aii claim: At a capped level this has in total increased from $15.3 million to $19.6 million, and at a pre-capped level has increased from $26.6 million to $33.4 million. In November 2001 the pre-capped figures had been as follows: Cloud - $6.4 million, Wind - $0.3 million, Shadow - $9.3 million, and Whisper - $10.6 million. In the trial claim, however, these pre-capped figures had become: Cloud - $17.4 million, Wind - $4.6 million, Shadow - $6.3 million, and Whisper - $5.2 million. Thus there had, in general, been a transfer from the latter two vessels to the former two, and in particular large increases in respect of those former two, the Cloud and the Wind. Nevertheless, the principle of the claim remained the same, namely a comparison, made as of six months after 9/11, of the actual income booked to be earned on those vessels in respect of their cruises down to the end of 2002, as against the income which, in the absence of 9/11, would have been expected to have been booked in respect of those cruises. Why then the differences between the November 2001 and the trial claims? To the extent that Silversea are limited to a claim of only $5 million across the fleet it would not ultimately matter. However, an understanding of the structure of the claim would afford greater confidence that the issues of construction which have to be determined are properly analysed. To the extent that there are reductions in the figures from November 2001 to trial, they may be explained by Silversea’s own pleading: thus at para 47 of their consolidated document they explain that losses on the remaining 2001 voyages had not been included on the basis that the number of last minute bookings for essentially the last quarter of the year were not likely to have been large; and that loss of income arising from passenger cancellations had been deducted on the ground that they were now claimed under the cover Ai claim. For the rest, however, I am not sure that I do understand just why the very large increases in the case of the Cloud and the Wind have occurred, other perhaps than that under the November claim the losses on the Wind were to be found under cover Ai rather than under Aii; and that with the passing of the time from November 2001 to the end of the six month period in March 2002 the actual bookings on the Cloud turned out to be even lower than had been anticipated in November.

As for claim B: this now amounted to $17 million capped at $5 million in the case of the Cloud and to $1.5 million, $1.9 million and $0.7 million in the case of the Wind, the Shadow and the Whisper respectively. Previously it had been $1.7 million, $21.3 million capped at $5 million, $2.9 and $0.8 million respectively. Again there has been a large shift from Wind to Cloud, possibly reflecting the fact that although the Wind was taken out of service, it was the Cloud’s itinerary rather than the Wind’s which was not performed. It would seem that the B claim is closely allied to the passenger cancellation claim. It is not clear whether the B claim represents the nominal value of credit certificates, or whether it truly reflects credits taken up and used on substitute voyages. It would only be the latter which would ultimately count towards a claim (see cover B’s “covers the cost to the Assured of issuing cruise credits), although given the value of such credits I would suppose that it would only be a minority of them which were not taken up.

I have already mentioned the fact that following the trial the judge held in his judgment that the original per diem claims had been abandoned and that he was only faced by the reformulated Ai claims based on passenger cancellations.

In their skeleton for this appeal, Silversea again argued their reformulated Ai claims based on such cancellations. They there submitted that the Ai cover was not based on loss of time, but on loss of income, for example – “If the Policy had intended to cover only loss of time, it could very easily have said so” (para 4.18 of their skeleton). There was however a very brief reference to the originally formulated per diem claims (at para 4.20.4) and a complaint that the judge had overlooked them.

During the argument of this appeal, however, as the difficulties of their reformulated claim bore in on Mr Flaux – in particular the difficulties of fitting a passenger cancellation claim into the structure of the Ai cover, at any rate in the absence of complete voyage cancellations, because of Ai’s language of interference “with the scheduled itinerary of the insured vessel” and also the Norwegian Plan’s reference to loss of time as well as the structure of the Sum Insured Schedule and of the policy’s deductibles in terms of days or the financial equivalent of days – so Mr Flaux began to seek to resurrect as an alternative the per diem claim based on cancelled voyages which had been Silversea’s original formulation. Indeed, in his reply Mr Flaux addressed himself solely to that original formulation and to Mr Swainston’s submission that it had been abandoned at the trial.

It is impossible and in any event undesirable to turn to the numerous issues of construction without at least some initial view as to the overall structure of the policy. If necessary that view can be further tested against the detailed arguments which arise on those various issues. I therefore turn to consider the question of overall structure.

In this respect the parties have differing views. On behalf of Silversea, Mr Julian Flaux QC submits that the essential difference between covers Ai and Aii is that Ai deals with loss of booked income and that Aii deals with loss of anticipated bookings: that is why Ai refers to “loss of income expected to be earned” while Aii refers to “loss of anticipated income expected to be earned” (emphasis added). It is therefore entirely appropriate that Silversea should claim for passenger cancellations under Ai and for loss of future bookings under Aii: for where a booking has been made the income is expected, whereas income from future bookings can only be regarded as anticipated.

On behalf of the underwriters, however, Mr Swainston submits that loss of both contracted reservations and future bookings comes within Aii as reflecting loss of market generally. The real distinction between Ai and Aii is that Ai is dealing with physical interference with the scheduled itinerary of a vessel leading to loss of time, whereas Aii is dealing with loss of market. If there is a difference between income expected to be earned and anticipated income, it is that once a voyage has begun, the fare is earned, whereas before it has begun it is only anticipated.

Mr Flaux rejoins that loss of time is not indispensable to cover under Ai and that even if it is it can be found where there is interference with the scheduled itinerary, a fortiori where a voyage is cancelled. Such voyage cancellations in response, for instance, to over-capacity caused by acts of war or armed conflict under the seventh peril listed under cover Ai amount to commercial, even if not physical, interference with the vessel’s scheduled itinerary. That is why in the case of the seventh peril the word “directly” is omitted from the phrase “which interfere with the scheduled itinerary of the insured vessel”, thus emphasising that commercial as well as direct physical interference is contemplated. It is also why the exclusion clause recognises “Deterioration of market and/or loss of market and/or lack of support for any scheduled cruise” as being relevant to a claim under Ai as well as Aii.

My preliminary view, to be tested as necessary by reference to arguments to be developed in the context of the detailed issues below, is that the critical distinction between Ai and Aii is that Ai is concerned with the immediate consequences of Ai insured perils upon the operations of the vessels and their current itineraries, whereas Aii is concerned with loss of custom or the less immediate consequences of Aii insured perils on future cruises. Thus I would refer to the following aspects of the language of the respective covers.

As for cover Ai: The first six (paragraphs of) perils listed under cover Ai are all concerned with occurrences of the kind which operate directly on vessels in the course of, or imminently about to commence, a particular cruise – viz, port closures, interruptions due to health problems, damage to repair facilities, blockage of any waterway, seizures, refusals of entry. Only the seventh peril, which lacks the adverb “directly” in qualification of the verb “interfere” and refers to acts of war etc may possibly, but not necessarily, have a wider impact. Moreover, the incorporation of the Norwegian Plan, the importance to its provisions of a loss of time calculation, the Sum Insured Schedule with its “per diem amount” column, and the deductibles provision expressed in terms of days, all support the concept that the Ai cover is concerned with loss caused by delay to the actual operations of the vessels, rather than to market losses expressed in terms of passenger cancellations or even, in extremis, the cancellation of whole voyages. Finally, the contrast between Ai and Aii cover is not so much between loss of “expected” income and loss of “anticipated” income, as between “loss of income expected to be earned by the operation of the insured vessels” and “loss of anticipated income expected to be earned on any future cruise”. The epithet “expected” describes income under both covers, but what is emphasised under cover Ai is “the operation” of the vessels, whereas what is emphasised under cover Aii is loss of anticipated income on future cruises.

As for cover Aii: This is concerned with loss of anticipated income on future cruises. The new reference to “anticipated” income emphasises the future. Unlike cover Ai, where loss of income is measured in terms of loss of time because the perils directly affect the operations of the individual vessels, under cover Aii the loss of income is measured directly in terms of income to be derived from “bookings”. However, since bookings provide only a gross measure of lost income, the language of the Aii cover expressly refers to a measure of the indemnity in terms of “Ascertained Net Loss”. Thus Aii covers “loss of anticipated income expected to be earned on any future cruise as detailed in the current Cruise Atlas” and the cover is further defined as “the Ascertained Net Loss resulting from a State Department Advisory…that negatively impacts the Assured’s bookings and/or necessitates a change to the scheduled cruise itinerary”. Since Aii is concerned with bookings on future cruises rather than delay to the vessels themselves, there is no express mention of “vessels” under it, as distinct from cover Ai where that word is repeatedly found (“operation of the insured vessels”, “loss due to the vessel being wholly or partially deprived of income”, “event which directly interferes with the scheduled itinerary of the insured vessel”, “acts of war…which interfere with the scheduled itinerary of the insured vessel”). Moreover, since cover Aii is concerned with loss of bookings on future cruises rather than the operations of individual vessels, the policy terms necessarily have to define both which future cruises are to be considered (namely all those “detailed in the current Cruise Atlas”, viz the 2001 and 2002 cruises) and over what time-scale the loss is to be measured (namely “subject to a maximum period per event of 6 months…”). Cover Aii also contemplates ascertained net loss resulting from a peril that “necessitates a change to the scheduled cruise itinerary”: but in the context that would seem to contemplate a loss of market sufficiently severe to require economically or commercially driven alterations to the cruises scheduled in the current Cruise Atlas.

As to cover B: Since this cover is subject to the same conditions as cover Ai, is paid for by a premium which is unified for both covers Ai and B, and is subject to deductibles which are combined for both such covers, it ought in theory to operate in the same way as cover Ai. My provisional view is that it does. It is concerned with “passengers” who have to be compensated for “loss of amenity” by the issue of “cruise credits” or “on-board credits for future cruises” where a cruise has been cancelled or interrupted by a peril falling within cover Ai. All of this to my mind suggests that the operations of a particular vessel have been affected so as to require passengers who are already on board (or perhaps are imminently due to board) and who have already paid for their cruise to be compensated by the issue of a credit. Such a credit may relate to a future cruise, that is to say, may consist in a credit against the cost of a future cruise, but in such a case the credit would seem to have to be an “on-board credit”, which seems to me to suggest that the passenger must already be on board the vessel whose operations have been affected by the peril concerned. It is common ground, however, in respect of the loss of time measure incorporated in the Norwegian Plan which forms part of the Ai cover provisions, and I would agree, that under cover B that measure is inapplicable: it is supplanted by a new measure of indemnity, namely “The reasonable cost of compensating passengers” by the issue of such cruise credits.

In sum my provisional view is that covers Ai and B relate to interference with the operations of a particular vessel, to be measured either by loss of time or by the cost of compensating passengers, whereas cover Aii is concerned with loss of market on the whole range of future cruises currently advertised and is measured on an ascertained net loss basis.

In such circumstances it would seem prima facie odd to think that a passenger cancellation or loss of market basis of claim could be applicable to covers Ai or B. A difficulty is nevertheless inherent in the fact that cover Ai expressly contemplates the possibility that interference with scheduled itineraries include voyage cancellations while cover Aii expressly covers not only loss of bookings but also changes to the scheduled voyage itinerary. There is in theory therefore an arguable possibility of an overlap in cover which is at the root of much of the dispute between the parties.

I turn therefore to the issues identified in paragraphs 32 and 33 above. They have been numbered in the order in which they were addressed in the submissions before us, an order which was driven by the contemplation that, as I have mentioned, the $5 million question was the most important dispute which separated the parties.

The Aii limit is “US $5,000,000 per event and in all…” The Ai limit is “As attached days per event and in all. Daily amount USD $ as attached.” The B limit is “US $5,000,000 per event.”

On behalf of Silversea, Mr Flaux submits that the $5 million limit applies to each vessel. He relies first and foremost on the general provision that “Each vessel to be a Separate Insurance”. He submits therefore that prima facie the limit applies to each vessel, for each vessel must be regarded as insured under a separate contract of insurance and each contract will contain the limit separately. The separateness of each such contract is emphasised by the fact that each vessel is owned by a different company, so that each contract has a different insured owner. He accepts that such a presumption is subject to contrary agreement, as in the case of the $200,000 premium for all four vessels: however, he says that there is nothing that amounts to a contrary agreement in relation to the limit. In particular, he contends that the wording “and in all”, which might have been posited as expressive of a fleet approach, is instead directed to capping losses due to more than one event. So also under cover Ai, where it is common ground that each vessel has a separate limit, the limit is expressed as “and in all” without aggregating losses of each vessel. The limit therefore aggregates losses caused by separate events rather than separate vessels. He was minded at trial to accept that the cover B limit of $5 million per event was a fleet limit, but on appeal he reserved his position on that.

On behalf of the underwriters, on the other hand, Mr Swainston submits that the $5 million limit is a fleet limit. He relies primarily on the wording “in the annual aggregate and in all” as being a definitive agreement that the limit is to operate in all circumstances, not only per event but also per vessel. He seeks to downplay the term in relation to separate insurance as merely intended as a protection to the assureds in case of some non-disclosure with respect to a single vessel. He relies on the across the fleet premium of $200,000 as indicative of an across the fleet limit. He contrasts the per vessel limits of cover Ai as being carefully worked out in terms of their size and earning capacity at different seasons of the year, and asks why, if cover Aii’s limit had been similarly intended to relate to each vessel, it was not similarly quantified to reflect the different sizes and earning capacities of each vessel. And he points out that Silversea’s ascertained net loss under the Aii cover can hardly be worked out vessel by vessel rather than across the fleet as a whole.

At the trial below the judge preferred the underwriters’ submissions (at paras 52/55 of his judgment). So do I, although perhaps with more hesitancy than the judge. In a policy which states that each vessel is a separate insurance, which provides for separate limits in respect of each vessel under its other main cover, and undermines the force of the expression “in all” by using it in connection with that per vessel limit under cover Ai, the point is not altogether free from difficulty. However, it is now accepted even by Silversea that the cover Aii premium provision is a fleet premium, which demonstrates that the policy may in certain respects operate across the fleet as a whole. Moreover, like the judge I am struck by the fact that cover Aii by its nature, and its wording, appears to produce a net ascertained loss across the fleet, and, as I have already remarked, the expression “vessel” is not even to be found within its language. Indeed, it is not easy to see how an ascertained net loss could be worked out on a vessel by vessel basis, although I appreciate that Silversea have sought to do so. As the judge said (at para 52):

Ultimately, the expression “in the annual aggregate and in all” is a strong pointer to a fleet-wide limit in a context where there is nothing to suggest a vessel limit and much to suggest a fleet-wide approach to the ascertainment of loss. Under cover Ai by contrast the full expression is “As attached days per event and in all”, which directly incorporates the Schedule’s separate vessel limits.

I am less sure that cover B’s limit is a fleet-wide limit, because cover B is closely associated with cover Ai, with its separate vessel limits in a context where cruise credits are intimately connected with a particular vessel’s cruise. Even then, however, a cruise credit could presumably be used in connection with a future cruise on a different vessel: but in such a case I would imagine that the cost of the credit would be debited against the owner of the vessel in respect of which the credit was issued. However, although Silversea’s concession below in respect of cover B was withdrawn, Mr Flaux did not advance any positive arguments in favour of a vessel by vessel approach to cover B, so that his case in this respect must remain uncertain. However, even if the cover B limit were a per vessel limit, and there is no issue which requires the determination of that question, I would not consider that that would ultimately affect the separate question under cover Aii.

On this first and for the parties highly important issue, I would dismiss Silversea’s appeal.

In the light of my answer to issue (i) there is no need to determine the underwriters’ alternative case of rectification under issue (ii). The judge was in the same position. It would seem, however, that he would have been willing to answer this issue, had it arisen, in favour of the underwriters and order rectification (at paras 56/65). Its subject-matter required him to consider in detail the evidence relating to the agreement of the premium for cover Aii and the circumstances which arose as to its invoicing and payment. He established, and it is now common ground, that there was no agreement for separate premiums for each vessel. It is unnecessary to repeat that exposition of the evidence. The judge concluded that there was no express discussion let alone agreement as to the number of limits under cover Aii, but nevertheless reasoned as follows (at para 64):

In the light of that indication, I think I should briefly state that, if I had concluded under issue (i) that the Aii limit was a vessel by vessel limit, I do not consider that there was any basis in the evidence or the judge’s findings to permit rectification. In effect the judge, correctly, found that the premium had always been expressly agreed in its lump sum form, whether one looks to the time of the first policy or to its later replacement. There never had been an express pre-contract agreement in any other form, nor had there ever been any discussion or agreement that the limit was to apply on a fleet wide basis. In such circumstances the judge’s finding that the parties had reached a pre-contract agreement on a fleet-wide limit which justified rectification amounted to nothing more than a finding of an implied agreement, contrary to the construction which, ex hypothesi, the judge had placed on the ultimate but identical policy wording. I do not believe that this is a legitimate basis for rectification, and the judge himself acknowledged that his conclusion was “at the extreme margin of what would ordinarily be achievable by way of rectification” (at para 64).

In my judgment the judge’s conclusion was inconsistent with the authority of this court in The Demetra K [2002] 2 Lloyd’s Rep 581. Lord Phillips of Worth Matravers MR there set out (at paras 22/24) the familiar principles which govern the proof of an entitlement to rectify, inter alia the need for an antecedent agreement which gives outward expression of a common intent, and convincing evidence sufficient to discharge the burden of proving a common mistake in translating the previous agreement into contractual form. Having considered the evidence in that case, Lord Phillips was prepared to allow that the negotiators had each assumed that their agreement on a certain matter would have a certain effect, but they had never discussed and agreed upon that effect. That, however, was not enough for rectification. Thus he said (at para 56):

On this contingent issue, therefore, Silversea have succeeded. In a sentence, although the evidence established that a single premium had been agreed, it did not follow that a single fleet limit had been agreed.

Cover Aii stated that it was “subject to a maximum period per event of 6 months from date that management within Silversea Cruises Ltd. shall determine and will so notify the Berkely Group accordingly”.

Silversea submit that the six month period is one for measuring the loss of bookings across the whole scheduled itinerary of cruises then being marketed, the so-called current Cruise Atlas, which went down to the end of 2002. The underwriters on the other hand submit that the six month period is one for defining the cruises on which loss of bookings is to be measured, viz those cruises which are due to commence during that period.

In this connection Mr Flaux emphasised that cover Aii is concerned with loss of income expected “on any future cruise as detailed in the current Cruise Atlas”, which he submits is inconsistent with the six month limitation. Mr Swainston on the other hand emphasised that it would be uncommercial and unfair, and inconsistent with the requirement of an ascertained net loss, to measure a fall-off in total bookings over a limited period, seeing that such a decline could be reversed immediately after the end of that period.

The judge was impressed by the latter argument. He said (at para 70):

My mind has fluctuated about this, but I have concluded that the underwriters’ submissions overstate the commercial importance of the possibility of subsequent recovery in load factors and underestimate the linguistic significance of the express cover given to loss of income “on any future cruise as detailed in the current Cruise Atlas”. The judge recognised but ignored the latter, primary, provision, holding that it had in effect been supplanted by the six month limit. But if that had been the parties’ intention, then the cover ought to have been defined in terms of cruises within the six month limit. Such a provision would have been easy to state. As the cover reads, however, the impact of an event is to be ascertained from its effect on the whole of the cruising programme, and not just six months of it, and the six months limit is rather a supplementary, not a supplanting, provision defining the period over which that impact is to be measured. It is perfectly true that an ascertained loss is an actual loss and that there is a possibility that a loss measured over an earlier period of six months might be compensated by an improvement over a subsequent period. It is equally possible, however, that any subsequent improvement, or any subsequent deterioration for that matter, may be due to matters entirely collateral to the event of the peril which has occurred. The parties have therefore agreed to allow only a six month period for determining and measuring the loss, accepting, as the judge said, that that is an arbitrary limit and that matters might well change thereafter. Just as a loss measured over a six month period of voyages is a real loss, but might be compensated by a better than expected market over the next period, so a loss measured over a six month evaluation of the market for the whole of the cruising programme down to the end of 2002 is a real loss of “anticipated income”, but might be compensated by better than expected conditions over the ensuing period.

I would therefore in this respect allow Silversea’s appeal, but I am doubtful that it will make much if any difference in the light of the $5 million overall limit on recovery under cover Aii.

It will be recalled that cover Aii requires not so much or not only acts of war, armed conflict or terrorist activities “actual or threatened” but “a State Department Advisory or similar warning” regarding such events. The loss of income covered is the loss resulting from such warnings. It will also be recalled that an exclusion clause excludes cover for any loss arising from –

The underwriters therefore argued below that Silversea’s claim for losses under cover Aii was uninsured because it was in respect of losses resulting directly from terrorist activities rather than directly from the US government warnings themselves. They sought to prove with the aid of an expert in management science, Dr Brian Gibbs, a visiting associate professor at MIT, that the post 9/11 deterioration of demand for Silversea’s cruises was only in small part (some 10/20%) due to the warnings and for the rest was caused primarily by the 9/11 events themselves. On their side Silversea deployed the evidence of a chartered clinical and occupational psychologist, Dr Reddy, to rebut any such allocation. The judge preferred the evidence of Dr Reddy, indeed he said that he did not even need it, so as to conclude (at para 68) that whereas

The judge found (ibid) that “it is impossible to divorce the effect of the warnings from the effect of the events which they so swiftly followed.” It would seem therefore that he found that the deterioration in Silversea’s market was inextricably caused directly both by the warnings and by the events themselves.

On this appeal the underwriters do not seek to go behind the judge’s rejection of their factual case on causation. They do, however, take a further point of law briefly referred to by the judge in this passing comment (at para 69):

Now on appeal at any rate they do. Mr Swainston submits that because of the exclusion only losses caused by government warnings are covered, not losses caused by the underlying events. Since, on the judge’s own findings, all the losses such as they may turn out to be were caused as much by the underlying events as by the warnings, it follows that the same losses would have taken place even in the absence of the warnings. It follows that on Dr Gibbs’ findings Silversea could recover for only 10/20% of their claim, whereas on the judge’s finding they could recover nothing.

It is common ground that the law is to be found encapsulated in this citation from Lord Phillips’ judgment in The Demetra K:

Both parties, however, submit that the application of these principles produces a result in their favour respectively. Mr Swainston submits that the 9/11 events themselves, because a direct cause of the losses different from the “insured event” under cover Aii, which has to be a warning, are excluded perils, and that losses caused by such perils are excluded losses. Mr Flaux, however, submits that the events of war or terrorism which lead to warnings are not excluded perils, but are perils covered elsewhere within the policy and are a necessary precondition, actual or threatened, of the warnings within cover Aii itself.

In my judgment Silversea are right about this. Cover Aii is premised on acts of war, armed conflict or terrorist activities, actual or threatened, provided, however, that they generate the relevant warnings about them. If they do, and those warnings cause loss of income as their direct result, there is cover. The underlying causes of the warnings are not excluded perils, it is simply that they are not covered under cover Aii as perils in themselves. Something extra is required. However, they are “an insured event” for the purpose of the contract as a whole. There is no intention under this policy to exclude loss directly caused by a warning concerning terrorist activities just because it can also be said that the loss was also directly and concurrently caused by the underlying terrorist activities themselves.

Since, however, the underwriters have undertaken to meet a cover Aii claim up to $5 million in any event, namely even if their defence under the exclusion clause were to be good, and since I have held that the overall limit under cover Aii is indeed $5 million, my decision on this point is, if my Lords agree with me, moot. If it were otherwise, the underwriters’ concession might have been a generous one.

I now come to cover Ai, which itself gives rise to a number of related issues.

There are in my judgment two grave difficulties in the path of Silversea’s primary claim under cover Ai, which has been formulated in terms of passenger cancellations or losses: one is the requirement of some interference with the scheduled itinerary of a vessel, and the other is that loss is to be measured in terms of loss of time.

As to the first, Mr Flaux has to rely not so much on the passenger losses themselves but on the cancellation of voyages due only to lack of demand (for the submission that any cancellations were due to lack of safety was rejected by the judge and has not been made the subject of appeal) as amounting to an interference with the vessels’ scheduled itinerary; as to the second, he either has to say that there is no need for loss of time at all, relying on the concluding words of the Norwegian Plan’s §16-6 “unless the circumstances clearly indicate otherwise”, or he has to say that a cancelled voyage represents a loss of time for the whole of that voyage. In either event, he is thrust back on to the original way in which Silversea formulated their claim, which was not in respect of passenger losses at all but in respect of the per diem losses of their vessels on cancelled cruises. I will deal specifically with those two ingredients under issues (vi) and (vii) below. For the present I ask the more general question whether cover Ai is in any event designed to indemnify Silversea against loss of market, or whether that is the province of cover Aii.

In this respect I have again considered the submissions of the parties set out at paragraphs 70ff above. There is however nothing in the detailed attention I have had to give to cover Aii under issues (i) to (iv) above that has caused me to doubt my provisional view that it is that cover which affords an indemnity for loss of market, whereas cover Ai is designed to compensate Silversea for interruptions in the operations of individual vessels in the course of or possibly in anticipation of their cruises.

Although it is in theory possible for separate covers to overlap, it would make no sense under a policy in which each cover has a separate regime for deductibles, the measurement of loss and limits on loss. Thus under cover Ai there prima facie has to be loss of time. The Ai deductibles are expressed in days. This confirms the relevance of the Norwegian Plan, which, as I have already remarked, says (§16-3) that “liability shall be calculated on the basis of the time during which the vessel has been deprived of income (loss of time)”, that loss of time is measured in days, hours and minutes (§16-4), and that each casualty shall be subject to a deductible period “which shall be reckoned from the beginning of the casualty and last until the loss of time…has reached the number of deductible days stated in the policy” (§16-7). The conceptual importance of loss of time to the measurement of loss under Chapter 16 of the Norwegian Plan, fully incorporated (apart from §16-1 itself) into cover Ai, can be seen from a perusal of that chapter’s full terms as set out in Annex 1. Cover Aii, however, has a deductible expressed in terms of “US $ 250,000 per occurrence”, not in terms of loss of time. Similarly cover Ai’s Sum Insured Schedule emphasises that loss is measured per diem in agreed liquidated amounts worked out separately for each vessel and each cruising season within the year, and also provides separate limits which reflect those separate per diem amounts: whereas cover Aii, as already discussed, provides for an ascertained net loss capped at $5 million.

All of this suggests that the two covers are dealing with separate kinds of loss: in essence loss of time in the case of vessels whose cruising and thus earning time has been interrupted or prevented, and loss of passenger bookings where government warnings have led to loss of market and/or the rescheduling of cruises for commercial as distinct from operational reasons.

It seems to me that the only possible point of overlap is where a cruise has had to be cancelled in advance because of the occurrence of a peril under either cover. Is the loss of income due to that cancellation a matter for recovery under cover Ai or cover Aii or possibly both? If the cancellation is due to the assured’s reasonable fear for the safety of the adventure, whether safety of the vessel or her passengers or both, then it seems to me to be at any rate feasible that the loss is covered by the seventh paragraph of perils found under cover Ai (“Acts of war, armed conflict, strikes, riots, and civil commotions which interfere with the scheduled itinerary of the insured vessel, whether actual or threatened”). If a claim based on such a cancellation were made, then an assured would of course not be able to rely on “terrorists” which is found mentioned as a peril only under the fifth paragraph of listed perils, but would have to bring himself within one of the other perils set out in the seventh paragraph. Such a cancellation would at least be an operational decision, and if such a decision were said not to be directly due to acts of war etc in the same way as some other immediate hindrance such as disease, blockage or refusal of entry might interfere with a vessel, then the assured could argue, as Mr Flaux has indeed submitted, that the absence of the word “directly” from the phrase “which interfere with the scheduled itinerary of the insured vessel” in the seventh paragraph (to be contrasted with that word’s presence in the equivalent phrase in the fifth paragraph) reflects just such an eventuality where the peril in question impacts with less than immediate directness on a scheduled cruise. That could well be the case with perils of the nature to be found listed in the seventh paragraph.

Suppose now that the assured could show that there were acts of war, and suppose there was also a government warning regarding such acts of war. Would the losses arising out of the cancellation of the cruise fall under both covers, or only one or the other? Since the whole cruise has been cancelled, the assured could argue that he has established a loss of time, at any rate if the vessel concerned could not be switched to some other cruise. In such a case, there could arguably be a claim under cover Ai or under cover Aii. Of course, the assured would not be entitled to be indemnified twice over, although he would possibly be entitled to show that there was some ascertained net loss over and above his recovery under cover Ai. I would, however, be inclined to suppose that such a claim, hypothetically based on the operational safety of the vessel and causing losses irrespective of the state of the cruising market, was to be brought under cover Ai rather than under cover Aii, which is concerned with loss of market. On the other hand, it is feasible that a cruise cancellation may have been caused both for reasons of operational safety and because the cruising market has suffered a decline which has made the cruise uneconomical. In such a case, I would be inclined to think that a claim under both covers is a real possibility, albeit one claim arises out of an “interference” with the vessel’s scheduled itinerary, and the other arises out of a “change” to it. “Interference” to my mind reflects the necessity under cover Ai for there to have been some operational impact upon the vessel’s itinerary (see also the word “Interruption” under the third paragraph relating to health hazards), whereas “change” reflects the more voluntary impact of a commercial decision taken in the light of the economics of the cruising market. I do not accept Mr Flaux’s submission that under this policy “interference” is apposite to describe a change of schedule adopted by Silversea to mitigate economic difficulties.

However in this case my hypothesis is not applicable, for on the findings at trial there have been no cruise cancellations due to safety, only (allegedly) due to market conditions. In such circumstances it seems to me that any claim properly belongs under cover Aii only. That is a fortiori the position where the claim is brought not in terms of loss of time but directly in terms of passenger cancellations.

In sum, even if Silversea are entitled to revert to their original per diem claims in respect of the cancellation of the voyages of the Wind, and even if they can establish that there have been acts of war or armed conflict within the meaning of the seventh paragraph, nevertheless, where such claims have to be premised on market as distinct from operational safety reasons, I would conclude that they have to be brought under cover Aii rather than cover Ai. That is still more clearly the case viewed in the context of their reformulated claim pressed at trial in respect of loss of passenger bookings.

Silversea invoke perils under the fifth and seventh paragraphs listed within cover Ai, and such paragraphs each contain a requirement that there be an interference with the scheduled itinerary of an insured vessel. It follows that the only claims which can possibly be made under cover Ai are those where the passenger losses are due to such interference. Since I believe that “interference” within cover Ai is a reference to operational and not market factors, and since the judge found that Silversea’s case on safety factors failed, I would conclude that on this ground alone Silversea’s reformulated claim does not belong and cannot be fitted under cover Ai. This is a fortiori the case where the claim in respect of loss of market is not tied in any way to any alteration in any vessel’s itinerary but only to loss of market (cf cover Aii’s “that negatively impacts the Assured’s bookings”). Mr Flaux was not willing formally to concede the latter point, but on the contrary submitted that all he need show, at least under the seventh paragraph of listed perils, was some economic consequence, even, in extremis, a single cancelled booking. That, however, was not only to water down the word “interfere” into the equivalent of “affect”, but also to give no point at all to the concept of “the scheduled itinerary of the insured vessel”.

In as much therefore as Silversea’s reformulated claim turns merely on showing loss of passenger bookings, the attempt to bring it within cover Ai is misconceived. To the extent, however, that Silversea seek to support their reformulated claim based on loss of passenger bookings by reference to underlying cancellation of voyages in turn caused by a loss of bookings or market, they can at least say that scheduled itineraries have been affected, but, in my judgment, cannot say that those itineraries have been interfered with, nor, fundamentally, are they able to bring such a claim of loss of market within the scope of cover Ai.

Nevertheless, Mr Flaux at this point of the analysis relied upon one further aspect of the policy which he submitted demonstrated a fundamental flaw in the opposing argument. He relied on the exclusion already discussed above in another context of –

The judge rejected this argument. He pointed out (at para 38 of his judgment) that the exclusion made good sense in the context of cover Aii – that is not in dispute. But, more controversially, he went on to suggest that it was of practical utility even in the context of cover Ai where, for instance, passenger cancellations which preceded a cruise subsequently cancelled on the ground of safety might lead to a claim under cover Ai. The judge considered such a claim would be misconceived because there would be “no actual impediment to performance of the itinerary as scheduled”: but the exclusion would perform the useful function of emphasising that lack of support for a scheduled cruise does not of itself trigger coverage.

I have some difficulty with that example. I think that a cruise properly and reasonably cancelled on safety grounds might well be said to give rise to a claim under cover Ai – but the claim would not be because of loss of market or support, but because of lack of safety. On the other hand, a cruise which might have to be cancelled because of lack of support, where that lack of support was directly due to a peril within cover Aii, would clearly be a valid claim within that cover.

I would prefer to answer Mr Flaux’s submission by questioning its premises that, because there are no words excluding cover Ai from the ambit of the exclusion, therefore it is intended to apply to both covers Ai and Aii; and that because it does so, therefore cover Ai must embrace loss of market, provided that is directly caused. In my judgment this is a weak argument. It might rather be said that the reason why two other exclusions are disapplied specifically from only cover Aii, is on the very ground that the loss of market cover there provided was intended to survive the exclusions: but that would make no sense if cover Ai was also intended to cover mere loss of market. Similarly, it could be said that the reason why there is no disapplication at all of the loss of market exclusion provided it is the direct result of an insured event, is because cover Aii is intended to give loss of market cover (provided it is the direct result of an insured event), whereas cover Ai is not intended to give loss of market cover at all. It would follow that the exclusion could not sensibly be disapplied from either cover, albeit for different reasons.

However, even if those alternative explanations were not regarded as powerful in themselves, the fact remains that Mr Flaux’s submission depends on a relatively long process of inference and in my judgment cannot compete with the more direct and stronger pointers to the respective ambits of covers Ai and Aii to be derived from their own wording. Moreover, it makes no sense to say, as Mr Flaux submits, that economic market losses come at any rate within the seventh paragraph of cover Ai’s listed perils because of the absence there of the word “directly”, and then find that such market losses are excluded by the exclusion in question on the very ground that they are not “a direct result” of an insured event.

I am not sure that by the end of the hearing Mr Flaux was seriously sustaining his submission that loss of time was unnecessary to Silversea’s reformulated Ai claim. On the contrary, he was anxious to stress those elements of the reformulation claim which involved the cancellation of cruises and thus might be argued to be premised on a loss of time. He also sought to renew Silversea’s original per diem formulation as a means of maintaining a loss of time based claim under cover Ai. I will revert to that original formulation below.

In my judgment I do not see how loss of time can be said to be unnecessary to a claim under cover Ai. The ramifications of chapter 16 of the Norwegian Plan, of the Sum Insured Schedule and of the limitation and deductible clauses have been referred to above. Nevertheless, Mr Flaux submits that the rewriting of the perils in §16-1 of the Norwegian Plan is an indication that the provision for a loss of time based calculation of lost income to be found in the following paragraphs of the Plan should also be jettisoned. He argues that the whole reason for re-writing the §16-1 perils was to tailor the insurance cover to a cruise line which, unlike an oil tanker, does not suffer loss of income by virtue of loss of time as such but by virtue of loss of passengers and the fare revenue which those passengers provide when they travel on cruises on its ships. However, I cannot accept that submission. If there had been any intention to make a more wholesale revision of the Norwegian Plan’s approach, then cover Ai would not have read “as per Chapt. 16 of the 1996 Norwegian Plan, but §16-1 to read as follows…”. Nor would the Sum Insured Schedule and the limitation and deductible clauses have been drafted in the way in which they are to be found in the policy. In any event, there is nothing fundamentally different in this respect between an oil tanker and a cruise liner: both earn their revenue essentially day by day by reason of their availability for undertaking cruises for which there is a consumer demand. An oil tanker’s revenue is taken in the form of freight or hire, voyage by voyage or under time charter, while the cruise liner’s revenue is primarily in the form of passenger fares. But if a voyage is interrupted or delayed or otherwise interfered with and time is lost, then that loss of time can be translated into a daily sum. The loss of time may or may not immediately impact on the fare revenues of the voyage in question, for those fares have in theory already become fully paid up on sailing: but they may impact on subsequent cruises, and loss of amenity on an interrupted cruise will bring with it losses in the form of cruise credits.

Similarly, the provision in the Norwegian Plan’s §16-6 “unless the circumstances clearly indicate otherwise”, whatever those words contemplate, cannot by itself undermine the strategic significance to cover Ai of its loss of time structure. Certainly the circumstances of Silversea’s reformulated cover Ai claim do not clearly indicate that loss of time is to be abandoned, especially when cover Aii has been plainly structured to deal with a claim based on loss of market.

A claim under cover Ai which does not reflect and rely on loss of time is therefore an improbable candidate. I am satisfied that Silversea’s reformulated claim based as it is on loss of passenger bookings and not on loss of time is an invalid one.

As Mr Flaux, pressed with the need to show both an interference with scheduled itineraries and loss of time, sought to reposition Silversea’s cover Ai claim to depend not so much on passenger losses as on a loss of time consequent on cancelled cruises, the question arose in the course of the hearing as to whether such an approach, which harked back to Silversea’s original per diem formulation, could be said to have survived (see paras 63 and 68 above).

The judge held that Silversea’s original claim had been abandoned (para 21 of his judgment). His draft judgment, when distributed to the parties’ legal representatives in advance of hand-down, had merely stated, in a single sentence, that –

At that time paragraph 21 of his judgment stopped less than half way through its ultimate full length, ending with the sentence which begins “Very substantial claims, in excess of US$24 million, are now advanced in respect of all four vessels based upon customer cancellations…”

That draft elicited a letter dated 17 March 2003 from Silversea’s counsel drawing the judge’s attention to a submission that the Wind lay-up claim, although reformulated, had not been abandoned. The letter concluded:

In the light of this letter the judge added a further section to his draft paragraph 21 so as to include the following:

The judge’s view that the original formulation had been abandoned was challenged in Silversea’s notice of appeal, but in a somewhat oblique manner, see paragraph 4.20.4 of the “Grounds of Appeal for insertion into section 7 of the Appellants’ Notice Form”:

Mr Swainston submitted that the judge was nevertheless right to say that the original per diem claim in respect of the Wind’s lay-up was abandoned, and he sought to support that submission by reference to passages in Mr Flaux’s opening and closing submissions (bundles CA2 at pages 24/25 and CA3 at pages 510/511) where Mr Flaux was emphasising to the judge that the reformulated claim is really premised on the cancellation of the Cloud’s voyages (out of safety concerns) and that the Wind’s lay-up “in itself is not relevant, it is an act of mitigation – what we are really talking about is the cancellation of the Silver Cloud voyages”.

I am not overly impressed by these citations, which I suspect in any event could be multiplied. Silversea’s primary case at trial, into which Mr Flaux no doubt put most, if not almost all, of his effort, was the reformulated claim. In pursuing that case, Mr Flaux went so far, and had to go so far, as even to disparage the basis of the original per diem claim, for he had to explain why a claim based directly on passenger losses rather than on a per diem basis was appropriate to cover Ai. In those circumstances I have every sympathy for the judge’s conclusion that Silversea’s original claim had simply been abandoned. Even if it had not been formally abandoned, it had been left as it were on the shelf, disregarded and disparaged.

Nevertheless, the fact remains that the original claim was not formally abandoned, but preserved, even down to closing skeletons. If the trial had been a final trial on all issues, that would not have assisted Silversea: they would either have to make their case, primary and/or secondary, or fail. If they put all of their effort into their primary case, they could not complain if they failed on their secondary case. However, the trial below was not a final trial, for quantum was left over for another occasion. Moreover, not only was quantum left over for another occasion, but, as the trial progressed, the lines between issues of principle and of quantum became blurred. This blurring even extended to so important, indeed fundamental, an issue as whether 9/11 had in fact caused so drastic a fall-off in the passenger market as to require Silversea to respond by reducing their cruising fleet to only three vessels. The underwriters said that that was a response to a cruising market which was already in down-turn, even before 9/11. There was a critical issue of causation here: but the parties agreed, and the judge sanctioned, leaving that issue over for a further trial (see paras 47/50 above). Moreover, built into either formulation of Silversea’s claim was this critical allegation that the Wind had been laid up because of 9/11’s impact on the market: see para 22.1(i) of Silversea’s consolidated document cited at paragraph 46 above.

In these circumstances, if I had concluded that Silversea’s original per diem claim would have been within the Ai cover, it would have been necessary to resolve the issue of its status in this litigation. As it is, that is probably unnecessary. I would however say that in my view the original claim was not abandoned. If therefore I were wrong about its incompetence as a claim under cover Ai, it would be necessary to decide how, if at all, that claim could or should be dealt with. In this connection I feel that this court has not heard enough about the full circumstances of the trial to say whether it would have been in the interests of justice or not to give Silversea any further leeway in preserving or promoting that original claim. In other circumstances, therefore, that question would probably have had to be remitted to the commercial court. Certainly, Mr Swainston has sought to submit that the original claim could not in any event proceed for substantive reasons which go beyond the points of construction which have had to be considered on this appeal. It is not clear to me, however, whether either for reasons of procedural fairness or because Silversea would in any event have failed to establish at trial necessary conditions for taking forward their original claim, it would be proper to say that, even if it were otherwise a potentially valid claim within the policy, the door had nevertheless shut on it.

It was Silversea’s case that, even if the rest of the perils included within cover Ai required physical interference with a vessel, nevertheless the seventh and last paragraph, because it left out the word “directly”, was wide enough to embrace purely commercial or economic decisions caused by 9/11. However, the only applicable perils within the seventh paragraph are acts of war or armed conflict. Silversea were therefore interested to demonstrate that 9/11 fell within those concepts.

The judge did not think that it was necessary to deal with this point, and in the light of my judgment above, the same position may be said to prevail on this appeal. Nevertheless, because the matter was extensively argued, I will briefly state the parties’ submissions in relation to the leading authorities and the relevant expert evidence in the case as to the political dimensions of 9/11.

It was common ground that the leading authorities on the peril of “war” (or “civil war”) as found in a commercial contract could be found in Kawasaki Kisen Kabushiki Kaisha of Kobe v. Bantham Steamship Company Limited [1939] 2 KB 544 (“KKKKK”) and Spinneys (1948) Ltd v. Royal Insurance Co Ltd [1980] 1 Lloyd’s Rep 406. In the former, Sir Wilfred Greene MR, in his judgment in this court, rejected any notion that the meaning of “war”, when found in a charterparty, was to depend on either the question whether war had been recognised by the Government, or on international law, or indeed on any technical meaning. The word had to be construed “in a common sense way”, in accordance with “the common sense of business men” (at 559). In the latter, a case concerning an insurance contract, Mustill J referred (at 426) to KKKKK and other authorities for the same propositions, viz that –

Mustill J then went on to consider whether there were any characteristics of an armed conflict which made it a war, albeit of an internal character, and to formulate at least three questions which he suggested were generally involved (at 429/430), namely (1) Can it be said that the conflict was between opposing “sides”? (2) What were the objectives of the “sides” and how did they set about achieving them? (3) What was the scale of the conflict, and of its effect on public order and on the life of the inhabitants?

We have heard conflicting submissions as to how such questions should be answered in this case. On either side the evidential underpinnings of such submissions were extracted from the unchallenged reports of Professor Wilkinson and Dr Gunaratna (see paragraph 36 above). Thus Mr Flaux emphasised that both Al Qaeda and the West, in particular the United States and its allies, could be identified as relevant “sides”; that they had opposing objectives, in the case of Al Qaeda the defeat of America, its forced withdrawal from Islamic lands and the recreation of an Islamic caliphate; and that the scale of 9/11, leading directly and swiftly to the undoubted war in Afghanistan, and more controversially to the “war on terror”, was unprecedented in any terms of mere terrorism. Mr Swainston, on the other hand, emphasised that the experts repeatedly referred to 9/11 as a terrorist attack and to Al Qaeda as a terrorist organisation; that there were no opposing armed forces or “sides”, only the hijackers and their victims, and that in any event one should be looking for sides which were “states” in an international law sense, whereas Al Qaeda did not qualify as a sovereign organisation or an organisation with sovereign as distinct from merely terrorist aspirations; and that Al Qaeda could not be said to be “at war” with the United States, and had no objective in terms of undermining the latter’s territorial integrity or of wresting power from it.

There were also conflicting submissions on the question of “armed conflict”, with Mr Flaux suggesting that the use of the airplanes as in effect flying bombs amounted to armed conflict, while Mr Swainston said that the attack was not “armed” in any conventional sense and that it is only with illegitimate hindsight that one might view the airplanes as potential weapons.

Since it is unnecessary for us to determine the question and since it is also a large and important issue, I would not propose to do more, in the context of acknowledging the cogency of the arguments of both sides, than to make the following brief comments. First, the contractual terms which need to be applied are “acts of war” and “armed conflict” rather than “war” itself. No authorities directly on those phrases were cited, but I could imagine that each of them might be broader than war itself in the sense that they could arise even in the absence of war. For instance, I wonder whether a casus belli, which it might be argued that 9/11 was, in as much as it led within a month to the invasion of Afghanistan, is a candidate for an act of war? In this connection, I also have in mind that the experts’ report stated that “Afghanistan under the Taliban was a state sponsor of Al Qaeda” (para 6.6). Secondly, in the light of authority, I am doubtful whether it is particularly helpful to think of war in its international law sense as armed conflict between sovereign states. I recognise of course that in the international law context the consequences of a decision as to war and peace are themselves important considerations: are terrorists to be treated as combatants or criminals? In the context of an insurance policy covering the world-wide operations of cruise vessels, however, I suspect that considerable weight has to be given to the concern of business men and their insurers over the scale and ramifications of the conflict. Thirdly, and for the same reason, the fact that everyone could agree that 9/11 was an example of a terrorist attack does not to my mind itself answer the question of whether it amounted to something more. The experts sought to put 9/11 in its context, pointing out that “The levels of organisation, co-ordination and planning required to execute these attacks were unprecedented” (para 2.1). They then used the rest of their report to provide that context, and did so by reference to separate headings dealing, for instance, with Al Qaeda and its structure and membership, its operation, its ideology and aims, its Taliban relationship, its worldwide scope, its strategy, the invasion of Afghanistan and the war on terror, the US and international response (“Both the US and international response to 9-11 was unprecedented because of the sheer scale of social and economic destruction caused by 9-11”, at para 11.1), and the threat posed by Al Qaeda (“13. Has the war on terror eradicated the threat posed by Al Qaeda?”). The report makes for rich and thought-provoking reading, but the submissions we heard barely scratched its surface.

For the reasons contained in this judgment, I would dismiss this appeal, save only in respect of issue (iii) above, where I would allow it.

I agree.

Rix L.J. has dealt so fully and lucidly with the complex issues that have arisen in this interesting appeal that I would ordinarily not do more than gratefully express my complete agreement with his judgment. I must, however, note one reservation.

For my part I do not believe that men of business, the underwriters and the insured, would have said as they watched those aircraft smash into the Twin Towers, “That’s an act of war!” They would have concluded, as the U.S. authorities described it in their Worldwide Cautions set out in paragraphs 37 and 38 of my Lord’s judgment that these were “terrorist actions from extremist groups … with links to Usama Bin Laden’s Al Qaeda organisation” and “retaliatory actions by terrorists … who harbor grievances against the United States”. To have declared “war on terror” was a rhetorical response serving only to emphasise that there was no identifiable “side” against whom the war could be waged. The war, a justified war, against Afghanistan truly was a war against that state fought to hunt down Bin Laden and his organisation. But it was not a war against them for war is not conducted against an individual and his wicked henchmen, however real a threat they pose to the security of a state.

To me an “armed conflict” has an air of continuity about it. Of course the threat from Al Qaeda remains and we must be constantly vigilant since they will undoubtedly strike again. But what are essentially random attacks do not carry for me, nor in my judgment for ordinary businessmen, the sense of present and persisting fighting between military groups which is the hallmark of armed conflict.

These are tentative views and I say no more because we are agreed that it is not necessary for us to rule upon these fascinating and important questions.