CL-2023-000064 - [2025] EWHC 1870 (Comm)

Colinvaux’s Law of Insurance (13^th Ed) at 11-357 opines that “The word ‘loss’ does not have a fixed meaning, and much will depend on its context”. The authors of that textbook refer first to Mitsubishi Electric UK Ltd v Royal London Insurance (UK) Ltd[1994] 2 Lloyd’s Rep 249, in which the Court of Appeal held that the use of a single defective component (a cementitious board) to make 94 toilet modules constituted a “single, albeit composite, head of loss”, such that the deductible (which applied to “each and every loss”) was only applied once, not 94 times. However, the relevant clause in that case referred to “each and every loss in respect of any component part which is defective in design plan specification materials or workmanship”, which wording clearly suggests the possibility that a single defective component part might cause a composite loss.

In Pennsylvania Co v Mumford [1920] 2 KB 537, the plaintiff was the custodian of securities. It was insured against theft, with a limit for “any one loss”. The Court of Appeal confirmed that there had been 41 losses, because the securities had been stolen by one employee on 41 separate occasions from four customers over seven years. It did not matter that they were all discovered on the same day (see, for example, p.547 per Warrington LJ).

In Glencore International AG v Alpina Insurance Co Ltd [2003] EWHC 2792 (Comm), one of a huge number of issues arising out of the collapse of Metro Trading concerned the operation of an $80m limit “any one loss” in the context of Metro Trading’s misappropriation of oil. Moore-Bick J held (at [291]-[304]) that there was a separate loss on each occasion when oil was misappropriated, even though this could be said to be “a single course of conduct repeated at frequent intervals over a period of time leading to what could be regarded as a single overall loss”(see [299]).

The learned judge explained his approach at [292]:

“There is no provision for aggregation in this policy other than whatever can be spelled out of the simple word “loss”. This is a policy against physical loss and damage to goods, so in the ordinary way a loss within the meaning of the policy occurs whenever the goods insured are damaged, destroyed or lost to the insured. Thus, several unrelated fires affecting goods in storage would give rise to several losses, as would several unrelated thefts. The position may be more complicated if several losses are related — as, for example, where an arsonist sets fire to two adjacent tanks in the course of a single attack — and no doubt a certain amount of common sense has to be applied when deciding how many losses have occurred in any given case. Thus, if thieves enter a warehouse containing bagged goods which they remove using a number of different vehicles, pausing from time to time to bring up a new vehicle, it is difficult to see how that could be regarded as more than one loss.”

He also explained (at [304]) that there might be a difference between the number of claims, or the number of causes of action, against Metro, and the number of losses:

“The insurance in this case is against physical loss or damage to the goods. The fact that a failure to redeliver the missing quantity may give rise to a single cause of action in conversion (if the claimant chooses to pursue a claim in that way) does not mean that there has been only one loss within the meaning of the policy. In my view each time MTI drew oil from the bulk and disposed of it without authority a loss occurred.”

His decision was followed by David Steel J in Dornoch Ltd v The Mauritian Union Assurance Co Ltd (No 2) [2007] EWHC 155 (Comm) which concerned a fidelity excess reinsurance with an “any one loss” deductible. An employee of the insured bank had, over 11 years, fraudulently transferred away client funds. It was held (at [31]-[34]) that each transfer of funds was a separate conscious act and hence “undoubtedly represented an individual loss which could have been the subject of a separate claim (on a policy providing cover for losses of this nature)” (see [33(2)]).

The result in Mitsubishi was different from that in Glencore or Dornoch, but, in each of the above cases, it might be said that the intellectual exercise is the same. One looks at the nature and terms of the cover to identify how an insured (or reinsured) loss is triggered. In Mitsubishi, it was using the defectively designed cementitious board to make various toilet modules. In Glencore, it was misappropriating the oil that triggered cover. If it had been property insurance, I agree with the Cs that a fire breaking out would have triggered the entitlement, and two fires breaking out in different locations would probably have been two losses, even if they happened on the same day, or were caused by the same arsonist.

In relation to the nature of and trigger for BI losses caused by COVID-19, in Stonegate Pub Co Ltd v MS Amlin Corporate Member Ltd[2022] EHWC 2548 (Comm), Butcher J dealt first with an issue as to the number of “triggers”. He described this concept of a “trigger” (which was not a word used in the policy with which he was concerned) as “a colloquial shorthand for the matter or matters which give rise to a right to claim under a policy” and observed that it could “be used to mean either the occurrence of insured perils, or the sustaining of loss as a result of the occurrence of insured perils” (see [59]). He equated the former meaning (at [61]) with deciding how many “Covered Event[s]” there had been, which was the phrase used in the definition of reduction of turnover: “the amount by which the Turnover during the Indemnity Period fell short of the Standard Turnover, less any costs normally payable out of the Turnover (excluding depreciation) as might cease or be reduced during the Indemnity Period as a consequence of the Covered Event” (see [20]). To my eye, “Covered Event” in the Amlin policy before Butcher J was being used in a similar way to the word “Damage” in the definition of the indemnity period in the Policy. See paragraph 44 above.

Using the word “trigger” in that sense, for the first kind of denial of access cover (enforced closure), he held (at [68]) that the “trigger”:

“…is the actual closure of all or part of an Insured Location under relevant compulsion or instruction. On this basis, the Policy is “triggered” in respect of each such closure, and the number of “triggers” is the number of Insured Locations so closed.”

In relation to the “per premises” argument, the conclusion above is to be contrasted with the view Butcher J took of the number of triggers in a second kind of denial of access cover which was before him, which required the actions or advices of a relevant agency to “have prevented or hindered the use of or access to Insured Locations”. In that context, he considered that the number of “triggers” was the number of actions or advices, not the number of premises. That was because (see [72]):

“The wording of the clause indicates that there will be a Covered Event if there is advice or actions from a relevant authority which prevents or hinders the use of or access to “Insured Locations”. While, in accordance with General Condition 7(ii), the plural will include singular, it is nevertheless the case that the clause provides that particular actions or advice by a relevant authority, though affecting more than one Insured Location, will constitute “Prevention of Access”. In those circumstances, the number of Covered Events under sub-clause (xii) should be regarded as the number of advices or actions rather than that number multiplied by the number of the Insured Locations to which those advices or actions related.”

That reference to the plural (“Insured Locations”) suggests to me that his answer might have been different if the clause had required the actions or advices to “have prevented or hindered the use of or access to an Insured Locations” (my changes underlined).

It might be noted that, in the related case of Greggs plc v Zurich Insurance plc [2022] EWHC 2545 (Comm), Butcher J explained (in a slightly different context) why it did not work for Zurich to say that the fact that there was a single “Business” for the purposes of calculating a reduction in turnover must mean that there would be a single loss. He observed (at [40]) that the aggregation provisions envisaged that there could be:

“…a plurality of losses which fall to be aggregated as one [aggregated loss]. Indeed, the logic of Zurich’s argument in this respect would appear to be that there was one loss…even if there were two wholly distinct Covered Events at different shops, because the disruption to each will be reflected in the overall Reduction in Turnover of the Insured’s Business. I regard that as clearly incorrect.”

In relation to the arguments about the effect of revised orders or regulations, Butcher J made clear in Stonegate that he did not accept that there were multiple triggers if there was a “reiteration, continuation or renewal of regulations which were, materially, to the same effect”. For the first type of denial of access “The “trigger” is the enforced closure, and in my view there will be one such “trigger” unless and until the Location opensand is then closed again” (see [69]). For the second type: “Steps taken or advice given by government or a relevant agency which merely repeated or renewed an existing prevention or hindrance of access would, in my view, form part of one set of “actions or advice”, and thus constitute one Covered Event” (see [73]). In Greggs, he made clear that the announcement of a measure and the regulations giving effect to it will usually constitute a single “Covered Event”, being actions or advice which caused the same prevention or hindrance (see [26]). In the other related case of Various Eateries Trading Ltd v Allianz Insurance plc [2022] EWHC 2549 (Comm), he observed (at [31]) that:

“The number of Covered Events must be judged by reference to the substance, not to the form or precise mode of promulgation or communication, of the relevant actions or advice. This will be achieved by looking at the groups of regulations, guidance or rules which brought about any particular prevention of access as being one Covered Event. I would not, for example, regard as sensible an approach by which lockdown 1 was regarded as having constituted separate Covered Events in respect of closure, stay at home/work from home and social distancing.”

It should be recognised that this part of Butcher J’s analysis of triggers and “Covered Events” preceded, and was not dependent upon, his discussion of the aggregation issues for each of these policy wordings (which primarily concerned the number of “occurrences”).

In that latter context, however, he did comment on the point in time at which the assessment as to whether there was one occurrence was to be treated as having been conducted. He held that it would be “the earliest time after the commencement of loss at which a reasonable person in the position of the insured would seek to decide whether there was one relevant occurrence” (see [94]).

I am not convinced that this timing issue is of particular importance to my assessment, since I am not concerned with aggregation in the sense of looking for a common cause, but rather with distinguishing between “one loss” and another loss. But I mention it because Mr Kramer relied upon Butcher J’s observation that one of the primary functions of business interruption insurance “is to provide the insured with funds during the interruption” (also [94]), and suggested that this meant that it was not necessary to wait until it was possible to determine how much loss, if any, had been suffered, before deciding whether there was one or more losses. That submission about timing seemed to me to miss its target, as I will explain, but I do accept that it would be surprising if it were necessary to be able finally to quantify each claim in order to know how many different losses had been suffered.

The U/Ws placed reliance upon Butcher J’s discussion in Greggs about whether announcements or measures “which simply continued existing restrictions or made trivial changes” would operate as separate single occurrences in the context of aggregation. He said (at [86]):

“I do not believe that it conforms to the parties’ intentions to have aggregation by reference to such matters, which effectively continued a status quo rather than marking any significant change to it. Nor would I consider that an informed observer would have regarded changes which simply reduced restrictions as being separate “single occurrences” for the purposes of the definition. They were such as would of their nature be expected to reduce losses not to lead to them and thus would not constitute the type of matter which would sensibly be regarded as a factor unifying different losses.”

That is a different question to asking whether an informed observer would have regarded changes which reduced restrictions as amounting to new trigger events or new “losses”. Of course, the answer to it might be the same.

In relation to the “per premises” argument, the Cs relied upon the decision of Cockerill J in Corbin & King(supra) where there was a limit expressed as “any one claim”. Cockerill J noted that the premises were in different locations, such that a danger or disturbance “would naturally give rise to two claims” and commented further that “The word “premises” points to each restaurant/ café and that distinction illuminates how a separation of interests may well operate – and that in turn points to separate limits. That then harmonises with the fact of different named insureds and the separate interests which underpin a composite policy” (see [239]). The composite policy point does not work in the same way in our case, but the examples which the learned judge considered neatly illustrated how the same danger might result in interruption at two different premises. Her expectation that these “would naturally give rise to two claims” chimes with my own view.

In International Entertainment Holdings v Allianz Insurance [2024] EWCA Civ 1281, the Court of Appeal confirmed the decision of Jacobs J at first instance, which Males LJ summarised (at [48]) as follows:

“…. the limit of £500,000 applied separately to each claim and that each closure of premises was a separate claim. To take as an example a policyholder with a theatre in Manchester and another in Oxford, the ability to claim for the closure of the Manchester theatre would depend upon proof of a relevant incident within the one-mile radius of that theatre, while the ability to claim for the closure of the Oxford theatre would likewise depend upon proof of a relevant incident within the one-mile radius of that theatre. That would be so regardless of whether the incidents in question were different in character (for example, an outbreak of Legionnaires’ disease in Manchester and a student riot in Oxford) or the same (two cases of Covid-19, one in Manchester and one in Oxford, each of which, adopting the analysis of the Supreme Court in FCA v Arch, was a separate incident). There was nothing in the clause to indicate that the limit of £500,000 was intended to operate on a per-insured basis.”

It might be noted that Males LJ was dismissive of the insurer’s argument that the fact that the clause provided cover in respect of “any claim resulting from interruption of or interference with the Business” was important because “each policyholder would have its own business, which might include some centralised costs, and that it did not make sense to speak of the business of the premises as distinct from the business of the policyholder” (see [49] – [50]). He did not find that argument persuasive.

Males LJ did indicate that it was important that the Allianz policy did not draw any distinction “…between those policyholders in the IEH group who own or operate only one venue and those who own or operate multiple venues. At the time when it was concluded, the policy did not even identify which subsidiary of IEH owned or operated which venue. So far as the policy was concerned, therefore, it was a matter of happenstance whether any particular subsidiary owned or operated more than one venue. To interpret the policy limit as applying separately to each policyholder rather than to each premises, when there is no clear wording to show that this was intended, would therefore be somewhat capricious” (see [51]).

To the extent that there is an overlap between a “claim” and a “loss” in this context, therefore, Corbin & King and International Entertainment Holdings might be said to favour a “per premises” approach, especially since some of the Cs have facilities which are geographically separate.

Finally, in Unipolsai Assicurazioni SpA v Covea Insurance [2024] EWCA Civ 1110, the Court of Appeal was concerned with a policy of property catastrophe XL reinsurance, pursuant to which an insurer was seeking to recover in respect of payments made to operators of nurseries for BI suffered as a result of COVID-19. One issue concerned when a BI loss occurred (or first occurred). Flaux C explained (at [146]-[147]) that:

“An “individual loss” first occurs when a covered peril strikes or affects insured premises or property and, when the covered peril which strikes the premises is the loss of the ability to use them (whether through damage to other property or premises or through a closure order as in the present instance) the individual loss occurs at the same point. It is immaterial for these purposes how the property or premises are affected and by what type of peril. The undisputed expert evidence was that market practice was and is to treat damage BI loss as occurring simultaneously with property damage and, like the judge at para 148(ii) I can see no basis for treating non-damage BI losses differently from damage BI losses…

147.

In other words, in all these cases, an “individual loss” only occurs once for the purposes of the Hours Clause, irrespective of how long the financial loss suffered continues for. It encompasses the entirety of the loss sustained by the original insured as a result of the relevant catastrophe striking or affecting the premises, irrespective of whether the relevant “individual loss” comprises physical damage losses, BI losses or both.”

He then explained the practical effect of this approach at [151]:

“…the interference with the business of each nursery which occurred on closure on 20 March 2020 was functionally equivalent to each nursery suffering physical damage on that day. The relevant “individual loss” occurred on that day and not day by day for every day that the business interruption continued. The answer to the example which Unipol posited of the vacationing business proprietors who did not suffer immediate interference and only suffered an individual loss when they would have reopened but for the closure order, was the one which Covéa gave: the business would still only suffer a single individual loss on that later date rather than on a day-by-day basis.”

This seems pretty clear to me. Whether one is dealing with damage or non-damage BI cover, a single individual loss can be said to have occurred once the business is interfered with by the closure, and that loss continues until the interference with the business comes to an end. The Court of Appeal appear to have considered that result to be consistent with the decisions of the English Courts concerning direct insurance (see [148]). I accept that both the policy wording and the context was different, but that decision does seem to me to offer important guidance as to the “shape” of a loss for the purposes of denial of access-type (or “non-damage”) BI cover.