Unipolsai Assicurazioni Spa v Covéa Insurance PLC

This appeal, with the permission of the judge, Foxton J, is from his order dated 28 February 2024 dismissing the appeal of the appellant reinsurer (“Unipol”) under section 69 of the Arbitration Act 1996 from an arbitration Award in favour of the respondent reinsured (“Covéa”) concerning the recovery of business interruption losses caused by Covid-19 under a Property Catastrophe Excess of Loss Reinsurance by which Unipol reinsured Covéa (“the Covéa Reinsurance”). The judge’s judgment dated 9 February 2024 also dealt with an appeal by another reinsurer, General Reinsurance AG (“Gen Re”), from another arbitration Award in favour of its reinsured Markel International Insurance Company Limited (“Markel”) which raised similar issues under that Property Catastrophe Excess of Loss Reinsurance. The judge also granted permission to appeal against his order in that appeal and this Court was due to hear the two appeals together, but that dispute was settled shortly before the appeal hearing.

Covéa provided cover to policyholders who ran children’s nurseries and childcare facilities, including under its standard NurseryCare Policy wording. As the judge said at [7] of his judgment (reflecting what the arbitration tribunal had found), that wording covered the wide miscellany of risks commonly found in commercial cover written by a property department, including cover for “business interruption caused by a peril other than physical damage to insured property” (“non-damage BI”).

The relevant terms of the Covéa Reinsurance are fully set out at Appendix 1 to the judge’s judgment. Of particular relevance to this appeal are the following:

“Condition 1 Reinsuring Condition

In consideration of the payment of the premium and subject to the terms and conditions of this Contract, the Reinsurers agree to indemnify the Reinsured up to the Limit(s) in excess of the Deductible(s) on account of each and every Loss Occurrence, which the Reinsured may sustain under the business specified in Class of Business, as stated in the Risk Details during the Period [of the Covéa Reinsurance] …

Condition 2 Definition of Loss Occurrence

1)

The term 'Loss Occurrence' shall mean all individual losses arising out of and directly occasioned by one catastrophe.

2)

The duration and extent of any 'Loss Occurrence' so defined shall be limited to:

…

vii)

168 consecutive hours for any Loss Occurrence of whatsoever nature which does not include individual loss or losses from any of the insured perils mentioned in any of the paragraphs (i), (ii), (iii) or (v) above

and no individual loss from whatever insured peril, which occurs outside these periods or areas, shall be included in that 'Loss Occurrence'.

4)

In all cases under this Condition 2 – Definition of Loss Occurrence …

3.

the Reinsured may choose the date and time when any such period of consecutive hours commences and the date and time when it ends, subject always to the maximum period of consecutive hours set out hereinbefore;”

In relation to the development of the Covid-19 pandemic, like the judge I will summarise the findings made by the arbitration tribunal (Sir Stephen Tomlinson, Michael Crane KC and Dominic Kendrick KC). The tribunal was provided with a detailed chronology which, as they found, revealed a “growing sense of crisis during the second half of February 2020 leading to an explosion of cases within the UK during the first half of March 2020…By 16 March, the date of the Prime Minister's broadcast and his advice to avoid non-essential social contact and travel, to work from home and to avoid all social venues, the number of cases along with the predicted rate of exponential increase in infections were threatening to overwhelm the NHS and to lead to many thousands of deaths…By March 2020, the Covid-19 pandemic had swiftly developed into a disaster engulfing the whole of the UK”.

The first death of someone in the UK who had tested positive for Covid-19 was recorded on 2 March 2020 and it was not until 5 March 2020 that it was made a “Notifiable Disease” in England. On 18 March 2020 SAGE concluded that the evidence supported school closures at a national level as soon as practicable to prevent NHS intensive care capacity being exceeded. On the same day, the Government instructed all schools, colleges and early years facilities in England to close with effect from 20 March 2020. That instruction was endorsed in law by the Health Protection (Coronavirus Restrictions) (England) Regulations on 26 March 2020.

Those restrictions were renewed for a further three weeks on 16 April 2020. On 1 June 2020, the phased reopening of schools, colleges and nurseries began in England. Then on 23 June 2020 the Prime Minister announced the lifting of all restrictions with effect from 4 July, effectively the end of the first lockdown.

By 8 June 2023, Covéa’s paid losses under nursery care policies amounted to £69.3 million plus £3.2 million in loss adjuster’s fees and Covéa sought indemnity for those losses under the Covéa Reinsurance. Unipol objected to payment essentially on two grounds which comprised the issues before the judge and before this Court:

Before the arbitration tribunal, Covéa’s principal case was that: “the outbreak of cases of Covid-19 in the UK in the period immediately preceding the closure of schools and nurseries on 20 March 2020 was a catastrophe”, alternatively that the various government orders or decisions constituted one catastrophe. The latter contention was introduced by an amendment following the decision of Butcher J as to what constituted an "occurrence" in the direct business interruption insurance policies he considered in Stonegate Pub Company Limited v MS Amlin [2022] EWHC 2548 (Comm).

The tribunal found that “if the idea of a 'sudden disaster'is inherent in the meaning of catastrophe” then “the exponential increase in Covid-19 infections in the UK during the first three weeks of March 2020 did amount to a disaster of sudden onset such as to qualify as a catastrophe” and that “the outbreak of Covid-19 disease in the UK during the few weeks preceding the schools and nurseries closure instruction ... amounted to a large-scale national disaster of sudden onset that may accurately be described as a catastrophe”. The tribunal held that the various instructions to close schools and nurseries issued by the UK and national governments “cannot be regarded as one or more catastrophes, at least not if they are to be viewed separately from the underlying pandemic to which they were a response”.

In relation to the Hours Clause, the tribunal rejected the argument that the reinsurance only responded to payments in respect of the closure of the insured’s premises during the stipulated period, finding that the reference to an “individual loss” in the clause meant “a loss sustained by an original insured which occurs as and when a covered peril strikes or affects insured premises or property”.

It is to be noted that in the Markel arbitration the tribunal reached the same conclusion that there was one catastrophe, but reached the opposite conclusion in relation to the Hours Clause, that it did only respond to payments in respect of the closure of the insured’s premises during the stipulated period.

The judge began his judgment by setting out the issues in the appeals and the principles as to the proper approach of the Court to a section 69 application. He then set out a summary of the background to the Award which I have replicated in the Factual Background section above.

In section B of his judgment he set out the applicable legal principles starting with the general principles as to the construction of the reinsurances, which were common ground and which he set out at [36]:

“i.

"The core principle [of construction] is that an insurance policy, like any other contract, must be interpreted objectively by asking what a reasonable person, with all the background knowledge which would reasonably have been available to the parties when they entered into the contract, would have understood the language of the contract to mean": FCA v Arch Insurance (UK) Ltd (The FCA Test Case) [2021] UKSC 1, at [47].

ii.

"Evidence about what the parties subjectively intended or understood the contract to mean is not relevant to the court's task" (ibid).

iii.

I was also referred to the summary of the general principles of construction in the judgment of Flaux LJ and Mr Justice Butcher in the Divisional Court decision in The FCA Test Case [2020] EWHC 2448 (Comm), [62]-[70]”.

He then set out the rival arguments as to the construction of aggregation clauses saying at [40] that he agreed with the approach of Butcher J in Stonegate at [84] that “in considering whether there has been a relevant 'occurrence' 'the matter is to be scrutinised from the point of view of an informed observer placed in the position of the insured'”.

The judge accepted at [44] that the history of a particular market wording (in particular the so-called LPO 98 wording) and the events which led to its introduction and modification do form part of the admissible factual matrix, at least where the parties to the contract are market participants and the materials are reasonably available to them. He referred to a number of cases where the Court had had regard to materials of this kind.

At [46] he noted that Unipol placed reliance on what they said was the market background to the use of the words “one catastrophe” in property catastrophe excess of loss reinsurance wordings as derived from Butler & Merkin’s Reinsurance Law (2022) and RJ Kiln’s Reinsurance in Practice (4^th edition 2001). The judge set out at [47] the origins of the LPO 98 wording as set out in Butler & Merkin and then what Mr Kiln (a leading Lloyd’s reinsurance underwriter from the 1960s to the 1980s) had said about how the wording developed. At [50] in particular the judge said:

“When discussing the background to the wording, Mr Kiln referred to issues which had arisen in the reinsurance market during the 1950s and 1960s as to whether losses arising from certain phenomena – for example a warm air front which generated a number of tornados, bush fires during a particularly dry summer, an exceptionally cold winter in the US which led to a greater level of motor claims and the cold winter in the UK in 1962/63 – could be aggregated for the purposes of claiming under property excess of loss reinsurance. Mr Kiln stated that in the revised wording drawn up against this background, the working party had used the words "occasioned by one catastrophe":

“because we felt it was more specific. It implied a violent happening which in itself caused damage. The word 'event' we felt might have applied to something which might have been the cause of a catastrophe rather than the catastrophe or disaster itself.””

The judge then set out at length in [51] what Butler & Merkin say about the subsequent history of the LPO 98 wording. At [52] he said that market debates discussed in these textbooks reflected one of the dividing lines in the type of reinsurance protection available. Kiln at 429-430identified the three options available to a reinsured:

“(a)

To take out a quota share reinsurance on his business. To do this, he pays a pro rata share of his premiums and receives a pro rata share of premiums and receives a pro rata share of all claims and expenses.

(b)

To take out an aggregate reinsurance to protect his business from a series of losses. This costs much less premium.

(c)

To take out an excess of loss contract to pay him if he suffers either a large individual loss or a series of losses arising out of some contingency. The contingency being a catastrophe, an accident an event or whatever. For this the premium he pays is much less than the quota share and much less than the aggregate premium (for a comparable limit and deductible). The Reassured has a choice and gets what he paid for.”

At [53] the judge said that: “The distinction between types two and three is often easier to identify conceptually than to demarcate in practice” and referred to the similar debate about asbestosis claims under catastrophe excess of loss reinsurances in the casualty market in the early 1980s which was the subject of two Lloyd’s White Papers in 1981, the second of which distinguished between “each and every loss” reinsurance contracts and “stop loss” or “aggregate loss” reinsurance contracts. The judge cites passages from that paper.

At [55] the judge said:

“I accept that the history of Article 6 of LPO 98 serves as an important reminder of the difference between a series of losses which can be linked at some level and which are catastrophic in their effect on the reinsured, and losses caused by a catastrophe properly so called. However, I am not persuaded that the materials before the court provide a basis for giving the word "catastrophe" in a property catastrophe excess of loss reinsurance any meaning other than that which it would bear on the application of ordinary principles of construction in the context in which it appears.”

He then set out his reasons for reaching that conclusion. The first was that the material was at a considerable remove, chronologically and textually, from the writing of the Covéa Reinsurance. The LPO 98 wording was produced in the late 1960s and thereafter amended twice. New Articles were produced by the LIRMA property catastrophe excess of loss clause which provided:

"a loss occurrence shall consist of all individual insured losses which are the direct and immediate result of the sudden violent physical operation of one and the same manifestations of the individual insured peril".

However, as the judge noted, a different wording again was used in the Covéa Reinsurance with no reference to “immediate result” or “sudden violent physical operation”. The judge considered that the connection between the wording in issue and the market debate in the 1960s was too tenuous for the materials to be used, not simply to identify in a broad sense the commercial purpose of a provision of this kind, but to ascribe textual limits not apparent from the ordinary meaning of the word in its contractual context to “catastrophe”.

His second reason was explained at [57]:

“Second, knowing the target at which the change in wording discussed in Butler & Merkin and Kiln was aimed – to address the argument that all losses from a severe winter could be aggregated for the purpose of collecting under an excess of loss reinsurance protection, or (per Reinsurance in Practice, 78) the argument that it was possible to aggregate by reference to "something which might have been the cause of a catastrophe rather than the catastrophe or disaster itself" – does not of itself tell you where the line of permissible aggregation is to be drawn in a very different context such as the present.” 

His third and final reason was that there had been significant changes in the content of books of direct insurance business in respect of which property catastrophe excess of loss reinsurance was purchased. As he said at [58]:

“The mere fact that, in the 1960s and 1970s, a reinsured's property account may not have included non-damage perils, with the result that a reinsurer providing (or indeed drafting) catastrophe excess of loss reinsurance for such an account would not have expected losses which impacted the cover to occur without physical damage to the original insured's property does not mean that the wording used in the reinsurance would not extend to such losses as a matter of its ordinary meaning. There is a distinction between the meaning of words in context, and their expected field of practical application from time-to-time, and market reinsurance wordings which are used for lengthy periods against a background of developments in the relevant book of business of the reinsured are, in a sense, "always speaking" in the manner of statutes (cf R v Ireland [1998] AC 147, 158-59).”

At [59] he noted that it was common ground that direct business written in an insurer’s property account now gives protection against business interruption even when it is not consequential upon damage to the insured’s property, so called “non-damage BI” cover, examples of which are denial or prevention of access cover of the kind in issue here and in Stonegate, including “notifiable disease” cover and loss of attraction cover. At [60] the judge said that there were a number of detailed findings about such non-damage extensions in the Covéa Award. He set out three findings of the tribunal:

“i.

“Non-property damage business interruption cover has been a common feature of many combined property/business interruption policies since about the beginning of the 21^st century and is now invariably written by property underwriters alongside the property damage risk, both as business interruption cover consequential upon damage to property and, under an extension, as cover for pure business interruption caused by perils other than damage to property”…

ii.

“By the time the Reinsurance was bound at the end of 2019, any competent and experienced catastrophe excess of loss underwriter reinsuring a UK property book would or should have known that the business reinsured might well include both business interruption cover consequential upon physical damage to an insured property and cover for interruption of the business carried on at an insured property from a peril other than physical damage to the property”…

iii.

“The unchallenged evidence … was that since the end of the last century it has become commonplace for the business written in property departments to include cover for business interruption from causes other than physical damage to property. Consequently, any experienced reinsurer underwriting the Reinsurance would know or ought to have known that the 'Class' of business written in Covéa's Property Department and classified as 'Household and Commercial' could, and probably would, include cover for non-property damage business interruption as well as for business interruption consequent upon property damage.””

At [61] the judge emphasised the importance of these findings given the terms of the Covéa Reinsurance:

“i.

The Class of business was defined by reference to that "written within the Reinsured's Property Department and classified as Household and Commercial and all business classified by the Reinsured as Contractors' All Risks and Engineering All Risks including Motor Own Damage".

ii.

Covéa was "the sole judge as to what is classified as 'Household' Business, 'Commercial' business and 'Contractors' All Risks and 'Engineering' All Risks business" (and there was no suggestion that the direct insurances which gave rise to its claims for indemnity were not properly so classified).

iii.

The premium payable to UnipolRe was to be calculated by reference to the "gross premiums of the Reinsured in respect of business coming within the Class (excluding Motor) written during the Period" (which would include any premium in respect of non-damage BI cover written in the relevant department).”

The judge went on at Section D of the judgment to address the rival arguments as to whether or not the individual losses in respect of which Covéa claim an indemnity arose out of and were directly occasioned by “one catastrophe”. At [66] he noted that neither party suggested that there was a settled and particular meaning of “catastrophe” in the reinsurance market. In those circumstances the judge accepted that a useful starting point in seeking to establish the meaning of the word as used in ordinary speech is the definitions given in dictionaries. He set out at [68] and [69] the definition in the full Oxford English Dictionary (“OED”) and the Shorter Oxford Dictionary (“SOED”) to which he had been referred.

At [70] the judge summarised the position on those definitions in these terms (I have omitted the specific references):

“I accept that both definitions embrace usages which refer to sudden events. However, they also show that the ordinary use of the word is not always so confined, with both dictionaries offering meanings which do not require "suddenness" (…including, but not being confined to, matters with the characteristic of suddenness). Many of the definitions emphasise the existence of a significant break with the position up to that point … and something which is seriously adverse in its nature or effects… The final usage offered in the SOED embraces all of these themes, and significantly offers "sudden or widespread or noteworthy" as alternatives. Further, the definitions offered include those appropriate to particular contexts (literary analysis or geology) which would have to be applied with care in other contexts.”

The judge then went on to consider in turn the three aspects of the meaning of the word “catastrophe” which Unipol argued applied here, but which it is said the tribunal erred in law in failing to recognise. The first was whether a catastrophe must be something which causes or can cause physical damage. At [72] the judge noted that Unipol did not seek to argue that in everyday usage “catastrophe” was confined in that way and he considered it right not to do so. Nor could it be argued that in the reinsurance market the term “catastrophe” was so understood. However Unipol did argue that such a requirement was inherent in the reference to “one catastrophe” in the Covéa Reinsurance.

The first point argued was set out by the judge at [74]:

“First, what was said to be the origin of the property catastrophe excess of loss class of business, which was said to go back to the San Francisco earthquake in 1906, and the origin of the LPO 98 wording following the physical damage claims brought following the severe winter of 1962/63.” 

However, the judge was not persuaded that the fact that non-damage BI does not appear to have been written when the LPO 98 wording was formulated in the 1960s confined the meaning of “catastrophe” in a market excess of loss reinsurance wording today. He agreed with Covéa that the established market practice by the time the reinsurance was written and the wording of the reinsurance provided strong support for the tribunal’s rejection of the supposed limitation on the nature of a catastrophe for the purposes of the reinsurance.

The second point argued was the description of the reinsurance as a property catastrophe excess of loss reinsurance. However, as the tribunal noted, claims for BI consequential upon loss of access to the insured’s property do involve interference with the insured’s premises so that cover for such losses were apt for inclusion within the scope of property catastrophe excess of loss reinsurance. The judge noted at [77] that this argument did not apply as forcefully to loss of attraction cover but said that even there the impact of the peril on the original insured’s property is an essential feature of any claim for indemnity.

The third point argued was that the perils specifically identified in the Hours Clause are of such a nature as to be capable of causing physical damage, the suggestion being that this gave rise to a contractual genus into which all catastrophes must fall. The tribunal rejected this argument and the judge agreed. At [78] he noted that, after listing various perils, (vii) of Condition 2 (2), the Hours Clause (set out at [3] above), referred to “any Loss Occurrence of whatsoever nature”. The judge agreed with the tribunal at [80] that the words “of whatsoever nature” were clearly intended to extend beyond unidentified members of the same “nature” as those specifically mentioned [i.e. at (i) to (vi)], the word “nature” being a powerful indicator in that regard.

At [82] the judge dealt with Unipol’s reliance on Condition 18 of the Covéa Reinsurance headed “Destruction by Civil authority”. This extended the reinsurance to include “direct loss and damage arising from the action or actions taken when complying with an order of a duly constituted Civil Authority…”. Unipol argued that this suggested that, absent such an extension, there would be no coverage for losses resulting from denial of access by government order. The Judge was not persuaded that this provision provides “any real insight” into the meaning of “catastrophe”. He said that the provision “appears to be directed to cases where the civil authority deliberately destroys property to protect against a peril” and one could well understand why the parties may have wanted to make it clear beyond doubt that loss and damage caused in this unusual way was covered.

The judge considered that the argument that the incorporation of a very specialist “extension” provided a basis for “reading down” the remainder of the reinsurance “is not a particularly attractive one” – the result would be that obtaining such an extension narrowed the remainder of cover than if the extension was not obtained. He said: “Many insurance and reinsurance contracts are assembled from a patchwork of pre-existing provisions drafted independently of each other, and that requires some care when seeking to determine the ambit of one "pre-packaged" provision from notionally "additional" cover provided by another.”

The second aspect of the meaning of “catastrophe” which Unipol contended the tribunal had erred in law in failing to recognise was that a “catastrophe” requires a sudden and violent event or happening. The judge noted at [88] that this argument rested on (i) statements in Butler & Merkin that catastrophes are to be “short, sharp and devastating” and in Kiln that what he had in mind was a “violent happening which in itself caused damage”; (ii) various of the dictionary definitions. The judge said at [90] that he had dealt with the historical materials [i.e. the LPO 98 wording and its evolution] earlier in the judgment and noted that it was only in the Butler & Merkin discussion of the LIRMA wording in which the word “sudden” appears. He said that he had addressed the dictionary definitions which offer meanings which involve sudden happenings but are not limited to such meanings.

In relation to the wording, at [91] he noted that it referred to riot, civil commotion and malicious damage which can be but need not necessarily be “sudden” in their inception, although he accepted that they will be violent. Riots and civil commotion can build up over time and while there are many cases where there is a point of “boiling over” the judge was not persuaded that this need always be the case, nor need they be short in their duration. The wording covered floods. The judge said that these can incept suddenly and violently, a “flash flood”, a dam bursting and so forth, but they can also build up over time. They can subsist for long periods. The judge also noted that the Hours Clause encompasses “collapse caused by weight of snow or water damage from burst pipes or melted snow”, which would once again “seem to encompass happenings which are not necessarily sudden in their inception or short in their duration, nor violent”.

The judge continued at [92]:

“identifying whether a happening is “sudden” will not always be a straightforward task, which suggests that some caution is required before treating this as an inherent but unspoken requirement for a catastrophe. Strong winds may build over time.”

The judge noted at [93] that the difficulties of importing a requirement of “suddenness” into the definition were also apparent in the approach taken in the two cases. The Markel tribunal had held that there was no requirement of suddenness, while the Covéa tribunal held that, if there was such a requirement, it was met by “the exponential increase in Covid-19 infections in the UK during the first three weeks of March 2020”, with 14 days elapsing between the first death of a person with Covid to the Prime Minister’s “stay at home” broadcast of 16 March 2020.

As for violence, the judge said at [94] that neither the dictionary definitions nor the words of the Covéa Reinsurance “suggest that a catastrophe must satisfy this requirement”. He accepted however at [95]:

“the radical discontinuity with what went before which is inherent in OED meaning 3(a) and SOED meaning 3… contemplates the ability to distinguish between the period when the catastrophe is in existence and when it is not… The more diffuse and extended the matter alleged to amount to a catastrophe is… the more difficult it may be to establish the coherent, particular and identifiable character which a catastrophe will have.”

The Full Federal Court of Australia had addressed that issue in Swiss Re International Se v LCA Marrickville Pty Limited [2021] FCA 1206 (answering the question: “was the outbreak of COVID-19 a ‘conflagration or other catastrophe’?”). At [355] of the judgment Derrington and Colvin JJ said:

“A number of parties submitted that it was not necessary that in order for an event to be characterised as a catastrophe it must involve an element of suddenness. That submission sits quite uncomfortably with the above dictionary definitions and those matters which might ordinarily be regarded as catastrophes: volcanic eruption, substantial explosion, earthquake, conflagration, tidal wave, a major deadly gas leak from a factory, cyclone, or hurricane. These examples support the necessity for a catastrophe to be sudden, or, at the very least, for it to have a commencement which is relatively certain in time and tend to eschew the inclusion of a state of affairs which emerges relatively slowly or progressively over time.” (the judge’s emphasis at [96].

The judge concluded on this part of Unipol’s argument at [97]:

“For these reasons, I reject the appellants’ argument that a catastrophe must necessarily be “sudden” in onset, or short in duration, or that it must be “violent”. Even if I had accepted that argument, it would not have provided a basis for challenging the Covéa Award in which the arbitral tribunal found that any requirement of “suddenness” was satisfied. There was no suggestion that this conclusion was “necessarily inconsistent” with the correct application of the relevant legal test, nor could that submission have realistically been advanced.”

The third aspect of the meaning of “catastrophe” which Unipol contended the tribunal had erred in law in failing to recognise was that a catastrophe has to be something which satisfies the unities of an “event”” as set out by Lord Mustill in Axa v Field [1996] 1 WLR 1026, 1035:

“In ordinary speech, an event is something which happens at a particular time, at a particular place, in a particular way... A cause is to my mind something altogether less constricted. It can be a continuing state of affairs; it can be the absence of something happening. Equally, the word “originating” was in my view consciously chosen to open up the widest possible search for a unifying factor in the history of the losses which it is sought to aggregate. To my mind the one expression has a much wider connotation than the other.”

Unipol relied on the use of the phrase “Loss Occurrence”, which has been held to have the same meaning as “Event” (Kuwait Airways Corporation v Kuwait Insurance Co SAK [1996] 1 Lloyd’s Rep 664, 686). The judge considered at [100] that “[t]his argument is in some ways both the easiest and most difficult of the issues raised on this part of the appeals”. The “easy answer” at [101] was that the Covéa Reinsurance does not use the words “Loss Occurrence” as a standalone term, but as a defined term whose meaning is set out in the “Hours Clause”. The judge could not accept “the argument of definitional determinism”, such that “the shorthand selected itself informs the meaning of the word beyond what appears in its associated definition.”

At [102] the judge noted that in the case of the Covéa Reinsurance “there are factors which point strongly away from anything but the most generous application of two of Lord Mustill’s three unities”. It contained a “two risk” warranty requiring the catastrophe to comprise losses covered by at least two different insurance policies before the Reinsurance could be engaged. The judge said that: “suggests the catastrophe has a potentially wide field of impact”. It also contemplates that something can have a duration exceeding 504 hours, and still be a catastrophe, and the only geographic limit imposed is a very broad one, for riot etc, of “one country” (and therefore the entirety of the UK). It was common ground that bush fires would constitute a catastrophe, and yet these can develop in a variety of locations over weeks, if not months. The judge also referred to what Lord Briggs said in The FCA Test Case at [323]: “a hurricane, a storm or a flood … may take place over a substantial period of time, and over an area which changes over time”.

At [103] the judge addressed the more difficult question which:

“is how to distinguish between a catastrophe properly so-called, which is an appropriate basis for aggregating individual losses when seeking indemnity under a property catastrophe excess of loss policy, and a series of discrete losses which share some common point of ancestry, but the adverse effects of which so far as a direct insurer is concerned are properly the subject of stop-loss protection (cf [52]-[53]). As Sir Jeremy Cooke observed in Simmonds v Gammell [2016] EWHC 2515 (Comm), [29], the "unities" are merely an aid to determining whether a series of losses involve such a degree of unity as to satisfy the contractual aggregation requirement.”

The judge concluded on this point at [104]:

“It is not necessary, for the purposes of disposing of these appeals, to provide a definition of catastrophe which can demarcate these distinct scenarios for all purposes, even assuming it is possible to do so. The answer is likely to be heavily dependent on the commercial and contractual context in which it arises. However, in the context under consideration here, I am satisfied of the following:

i.

The catastrophe must be something capable of directly causing individual losses, because that is what both "Hours Clauses" require. That requirement of itself is likely in most if not all foreseeable scenarios to exclude attempts to aggregate by reference to what are often described in aggregation disputes as "states of affairs".

ii.

The catastrophe must be something which, in the context of terms of the Reinsurances in which the term appears, can fairly be regarded as a coherent, particular and readily identifiable happening, with an existence, identity and "catastrophic character" which arise from more than the mere fact that it causes substantial losses.

iii.

To that extent, it ought to be possible, in a broad sense, to identify when the catastrophe comes into existence and ceases to be, even if an attempt at a precise temporal delineation would offer scope for legitimate debate and dispute.

iv.

A catastrophe will involve an adverse change on a significant scale from that which preceded it.”

The judge’s overall conclusion that there was a “catastrophe” was set out at [105]-[106]:

“105.

The Covéa tribunal recorded the "explosion of cases" from the second half of February to the middle of March, the Prime Minister's broadcast and the closure order. In the "Award and Disposition" they found that:

"the outbreak of Covid-19 in the United Kingdom, reflected in an exponential increase in the number of infections during a period up to and including 18 March 2020, was a 'catastrophe' within the meaning of Condition 2(1)."

106.

Having rejected UnipolRe's legal arguments at [72] to [102] above, there is no basis on which it can be said that this answer is "necessarily inconsistent" with the proper interpretation of the word "catastrophe" in the Covéa Reinsurance, indeed quite the contrary:

i.

There has been no suggestion that the catastrophe so identified did not directly occasion the original losses in respect of which indemnity is sought…

ii.

In the context of the Covéa Reinsurance, the "outbreak" described by the Covéa tribunal can fairly be regarded as a coherent and discrete happening, with an existence, identity and "catastrophic character" which arise independently of the fact that it causes substantial losses. As the Covéa tribunal noted, "during this relatively short period, the Covid-19 outbreak assumed a certain coherence in its development and effect and gave rise to a profound subversion of the order of life within the UK".

iii.

The Covéa tribunal identified the relatively short period within which the catastrophe came into existence.

iv.

The Covéa tribunal noted the (undisputed) wholesale disruption to our national life which the outbreak occasioned.”

Accordingly, the appeal against the conclusion of the Covéa Award that there had been a catastrophe for the purpose of the Reinsurance was dismissed by the judge.

At Section E of the judgment, the judge addressed the Hours Clause. At [111] he summarised the argument of both Reinsurers as being that, even if there had been a catastrophe for the purposes of the Reinsurance, only BI during the 168 hours stipulated by the relevant section of the “Hours Clause” could be relied upon for the purposes of seeking an indemnity under the relevant Reinsurance.

The judge considered first the position where business interruption losses are suffered as a result of damage to insured property. At [113] he recorded that it was common ground in both arbitrations that, in a conventional scenario, where an insured peril damages commercial property such as a hotel, which then endures a period of interruption to its business while it is repaired, for the purposes of a property catastrophe excess of loss reinsurance, the entire loss which the hotel owner recovers under the direct insurance is treated, for the purposes of any temporal limitations in the reinsurance, as having occurred on the day of the property damage.

The judge noted at [115] that, in the arbitration, Unipol had argued that this was “driven by pragmatic considerations and wrong in principle”, but that in such a case, there was undoubtedly physical damage when the peril hit, on which the business interruption was “parasitic”. At [118] the judge recorded that both reinsurers sought to identify a principled distinction between the market’s approach to “damage business interruption” and what they contended was required for “non-damage business interruption”. The judge then dealt at [119] to [128] with the arguments advanced by Mr Kendrick KC on behalf of Gen Re in the Markel arbitration, concluding that the distinction he sought to draw was significantly overstated and, on the contrary, there were strong similarities between damage business interruption and non-damage business interruption.

At [129] the judge recorded the submission of Mr Aidan Christie KC on behalf of Unipol that different insured perils are involved when considering claims for damage and non-damage business interruption. He submitted that in cases of damage business interruption the insured peril is the damage to the property, relying on [215] of the judgment in the FCA Test Case. At [130] the judge said that in considering this submission it was helpful to set [215] in its surrounding context and quoted [214] to [216] of the judgment.

At [131] the judge rejected Unipol’s argument in these terms:

“I accept that a necessary element in a claim for damage business interruption is that the business interruption results from physical damage itself caused by an insured peril. However, that is also true of some forms of non-damage business interruption claims (see [120]) and it is not of itself particularly informative. If it matters (and I am not persuaded that this does provide a principled basis for any difference in treatment on its own, in any event), I do not read Arch, [215] as suggesting that damage to property is the insured peril to which damage business interruption cover responds. At [215], the Supreme Court is equating the interruption to the policyholder's activities in a non-damage business interruption claim with the "destruction of or physical damage" in the Court's three stage "insured peril > proximate cause > physical damage" sequence. The implication of the Supreme Court's analysis is that the proper sequence for non-damage business interruption of the kind they were considering was “insured peril > proximate cause > business interruption…”

The judge then considered an argument on the Hours Clause which because of differences in the wording was run in the Markel arbitration but not run in the Covéa arbitration before turning to the question when does an individual loss occur for the purposes of the Hours Clause. He recorded at [135]-[136] the conflicting views of the arbitration tribunals:

“135.

The Covéa tribunal concluded that an “individual loss … occurs” for the purpose of the “Hours Clause” when the nurseries were closed on 20 March 2020, even though the business interruption continued until the nurseries were allowed to re-open when the first lockdown restrictions were lifted…, that being when “the insured first sustains indemnifiable business interruption loss within a nominated 168 hour period”, with loss which the insured continues to sustain afterwards being aggregated with the loss sustained during the 168 hour period.

136.

By contrast the Markel tribunal took the view that the original insured’s business interruption losses occur “day by day”… and that only those losses which occurred (on that construction) during the 168 hour period can be recovered...”

The judge said that before considering these competing analyses it was helpful to set out how business interruption losses are assessed under direct insurance policies. He did this at [138]-[142]:

“138.

Business interruption insurance provides cover during the period which it takes the business to return to normal trading (i.e. the level of trading which would have prevailed but for the operation of the insured peril), subject to a “Maximum Indemnity Period” which will provide a cut-off. There was a three-month cut-off in the Covéa NurseryCare Policy (Covéa Award, [92]). As UnipolRe and General Reinsurance noted when seeking to address the argument, there can be substantially longer Maximum Indemnity Periods, for example 36 months is not uncommon.”

[The judge set out an extract from Riley on Business Interruption Insurance at [139]]

140.

Further, within the relevant indemnity period (i.e. the actual recovery period, or, if shorter, the Maximum Indemnity period), the amount of the indemnity is not calculated on a “day-by-day” basis at the direct policy level, but across the period, with claims for increased cost of working, and credits for saved expenses…”

[The judge set out a hypothetical calculation from Riley at [140]].

“141.

There is usually provision for adjustment of the trend of the business which is the subject of the Business Interruption claim…” [The judge then set out the explanation by the Supreme Court in The FCA Test Case at [253]-[254].

142.

It will be apparent from the above [the passage from The FCA Test Case] that the amount paid to settle an individual business interruption loss can reflect a combination of credits and debits over a certain period, and that there may be considerable variation over time before you arrive at the final amount. This is very far-removed from a “day by day” calculation which UnipolRe’s and General Reinsurance’s arguments appear to assume. It is also clear that the assessment of a Business Interruption loss at the direct insurance level involves looking at the net effect over a particular period, not the aggregation of a series of daily losses (which answers Mr Christie KC’s point that Covéa’s case involves a “double aggregation”).”

The judge then considered other provisions of the Reinsurances he determined were of relevance when considering the meaning of the words “individual loss which occurs”. In the case of the Covéa Reinsurance, there were two provisions concerned with the timing of losses to which he referred in [144]:

“i.

The Reinstatement Provision provided “Losses hereunder are applied chronologically by date of loss.” In the case of pure business interruption losses, it is difficult to see how this provision is to operate if there are separate losses day by day (or even hour by hour).

ii.

The Extended Expiration provision addresses the position where the Reinsurance expires or terminates “while a Loss Occurrence is in progress”, providing that in such a scenario, UnipolRe are liable “as if the entire loss or damage had occurred prior to the expiration or termination of this Contract provided that no part of that Loss Occurrence is claimed against any renewal or replacement of this Contract.” … it would be surprising if the Reinsurances responded to physical damage suffered after the expiration date and consequential business interruption caused by a hurricane which had commenced before and continued to operate after that date, but not business interruption loss which continued after the expiration date, caused by a result of denial of access resulting from physical damage to neighbouring property which was complete before the expiration date.”

At [145] he considered other clauses in the Covéa Reinsurance which addressed how the quantification or assessment of losses at the insurance level impacts on the reinsurance:

“…the Ultimate Net Loss clause, with the allocation of loss adjustment expenses, litigation costs and the application of salvage and recoveries, and the "follow the settlements" clause. On UnipolRe and General Reinsurance's case, business interruption losses and associated expenses and credits paid by the reinsured have to be unpicked at the reinsurance level to distinguish between expenses and credits referable to interruption during the relevant "Hours" period, and that referable to any subsequent period. While this is an aspect of a more general issue where a settlement or loss at the reinsured level reflects both losses which are reinsured and those which are not, it does present that difficulty in a particularly acute form.”

The judge set out his analysis of this issue in the case of the Covéa Reinsurance at [147]-[148]:

“147.

It was common ground in both appeals that the references in the two “Hours Clauses” to “individual losses” mean the loss sustained by the original insured. That is significant, because it points the enquiry in the direction of the direct insurance. Further, as the Covéa tribunal noted, the “Hours Clause” defines “the extent and duration of a ‘Loss Occurrence’”… not the duration of an “individual loss”...

148.

As the Covéa tribunal found…, and as was common ground before the Markel arbitration, when considering damage business interruption, the individual loss occurs when the insured peril damaged the insured premises. I agree with the Covéa tribunal that this supports an analysis which treats an individual loss as occurring “when a covered peril strikes or affects insured premises or property” …, and that when the insured peril which strikes the premises is the loss of the ability to use it (whether through damage to other property or premises or a closure order), the individual loss occurs at the same point.

i.

That reflects the position at the direct level, to which the words “individual loss” naturally direct attention and, as the Covéa tribunal noted at Covéa Award, [97], “there is nothing in the Reinsurance wording to support an apportionment of an ‘individual loss’”.

ii.

Not only am I not persuaded that there is any sufficient basis to distinguish between the treatment of damage business interruption, business interruption following damage to other property owned by another and “pure” business interruption losses in this regard, but there are obvious parallels between the impairment of the rights of those entitled to property resulting from damage and that resulting from the inability to use the property: [120].

iii.

That analysis is consistent with the approach of Mr Justice Butcher in Stonegate and Various Eateries and the Court of Appeal in the latter case and the Supreme Court in FCA v Arch ...

iv.

This construction better coheres with the provisions dealing with the timing of the individual losses in the Reinsurances…

v.

It avoids the uncommercial consequences of the “day-by-day” construction as outlined at [138]-[142] and [145]-[146] above and [150]-[151] below, and in that respect derives some limited support from the provisions of the Reinsurances quoted at [145]-[146] above.”

At [150]-[151] the judge set out a number of the practical difficulties identified by Unipol with the construction which appealed to the Covéa tribunal and the judge and rejected them:

“i.

First, he posited the example of a closure order being made in respect of premises where the occupier was on holiday at the time when the closure order took effect (e.g. a cotton factory during "Whit week"), such that there was no impact on the business until the factory reopened.

ii.

However, the interference with the owner's right to use the factory occurs when the order comes into effect, whatever use the owner wishes to make of the factory at that point in time. Lord Halsbury's celebrated observation in The Mediana [1900] AC 113, 117 is in point:

"Supposing a person took away a chair out of my room and kept it for twelve months, could anybody say you had a right to diminish the damages by shewing that I did not usually sit in that chair, or that there were plenty of other chairs in the room? The proposition so nakedly stated appears to me to be absurd."

This was effectively the point made by the Covéa tribunal.

iii.

The "Whit week" closure would no doubt be factored into the calculation of loss during the indemnity period (as would the reverse position, where the first week involved the most profitable contract of the year).

iv.

The same issue could equally arise when there is physical damage during a period of holiday.

v.

Indeed, this argument rather points to the difficulties of UnipolRe's construction, confining cover to the position over 7 days, when profits, costs and savings are likely to be "lumpy" in their effect, rather than play themselves out on a linear basis over time. What is to happen, for example, where the insured premises experience an initial saving in costs over the first seven days (for example because the bulk of children were not due to return for another week after the date the closure order came into force), but a significant net loss over the indemnity period as a whole? On UnipolRe's approach, it is not clear whether this precludes an indemnity under the Reinsurance in respect of that claim, nor what the position would be in the reverse cases, where the adverse effects on the insured business are "front-loaded".

vi.

This is also true of the provisions in the Covéa and Markel Reinsurances allowing a "Loss Occurrence" or catastrophe which is "greater than the above periods" to be divided into two or more "Loss Occurrences" or "Events": it would involve "slicing and dicing" what, at the direct insurance level, is a net loss arrived at taking account of debits and credits over the indemnity period into account, so as to place constituent parts of that calculation into separate periods.

151.

Second, he posited the example of a (for example) retail premises undergoing repairs scheduled to last many months when a closure order took effect which allowed workers to attend at the premises, but not customers. He suggested that no business interruption "loss occurrence" could occur until the premises re-opened, and that it would make no sense for there to be a different result if "the direct insured's business was interrupted, however fleetingly, when the restriction on access was first imposed." However, I do not accept the correctness of Mr Christie KC's premise, to which the points made in the preceding paragraph are equally apposite.”

The judge concluded at [153(i)] that Unipol’s appeal against the conclusion of the Covéa tribunal on the issue of “one catastrophe” and on the operation of the Hours Clause in the Covéa Reinsurance should be dismissed.

Unipol appeal against the judge’s order on two grounds which can be summarised as follows:

In this section of the judgment I will summarise the written and oral submissions of the parties. Where authorities were cited, I will set out the passages in question, avoiding repetition in the later Discussion section of the judgment.

Before turning to its analysis of the two issues, Unipol submitted that the judge had made two fundamental errors which underpinned his analysis of both issues. The first error was said to be that the judge conflated the nature of the cover under a Property Catastrophe XL reinsurance with the cover provided under a stop loss/aggregate XL reinsurance. Catastrophe XL reinsurance is qualitatively different from other forms of reinsurance in that it responds only to causative events. It is triggered by events that cause loss, not the losses themselves. By contrast, quota share and stop loss/aggregate treaties are essentially responding to losses (and are not concerned with causation). Unipol contended that the judge had failed to recognise this distinction in his judgment.

The second fundamental error was said to be that the judge wrongly assumed that the terms of the direct policies could be read across into the Reinsurance. It is axiomatic that the Reinsurance should not be construed in light of the underlying policies: see Axa v Field. XL reinsurance is “not in any real sense back-to-back” with the direct insurances. The Judge referred to this principle at [126] but failed to apply it. His and the arbitration tribunal’s analysis of what was a “catastrophe” for the purposes of the Covéa Reinsurance relied heavily on the evolving types of cover written by direct property underwriters since the 1990s. Mr Christie KC and Ms Noakes submitted that this was a failure of logic: the nature of a “catastrophe” in a reinsurance contract cannot be conditioned by the types of cover written at the direct level.

The same error was said to infect the judge’s analysis of the Hours Clause. His reasons for rejecting the proposition that business interruption losses occur day-by-day included his reliance on the methodology for calculating and assessing business interruption losses at the direct level: [137]-[142] of the judgment. Unipol contended that, in his discussion at [121]-[126] (in the case of the Markel Reinsurance) and [130]-[132] of the Covéa Reinsurance, the judge appears to have relied upon the terms of the underlying policies to support the proposition that an “individual loss” occurs when a covered peril strikes insured premises.

In relation to the first ground of appeal, Unipol submitted that because Condition 2(1) in the Hours Clause expressly provides: “The term ‘Loss Occurrence’ shall mean all individual losses arising out of and directly occasioned by one catastrophe” a “catastrophe” under this reinsurance can only be an event or a species of event. It is only an event which is capable of “directly occasion[ing]” individual losses. Mr Christie KC and Ms Noakes submitted that the words “directly occasioned” imported a strong causal link which was only satisfied by a species of event, not a state of affairs. They submitted that the judge had accepted this in his conclusions at [104] (cited at [44] above).

The tribunal had purported to identify the “catastrophe” as “the outbreak of Covid-19 in the UK reflected in an exponential increase in the number of infections during a period up to and including 18 March 2020”. Mr Christie KC and Ms Noakes submitted that there had been no evidence as to the rate of infection so as to justify a finding that there was an explosion of new cases or an exponential increase. In any event, this putative “catastrophe” could never, on the judge’s own analysis, amount to a catastrophe for the purposes of the reinsurance as it was not a “happening” much less a coherent, particular and readily identifiable one. They submitted that the outbreak of Covid-19 was not an event or an occurrence but a state of affairs and more importantly, it was part of a global pandemic which was itself a state of affairs which began in late 2019 and is still going on. The outbreak in England could not be isolated from that overall state of affairs as the tribunal and the judge had sought to do artificially by reference to a vague and undefined “explosion” or “exponential increase” in infections. This was not a “catastrophe”, but a description of a state of affairs.

In support of Unipol’s case that wording such as that in Condition 2(1) “arising out of” or “directly occasioned” by one catastrophe is only satisfied if the “catastrophe” is an event, Mr Christie KC and Ms Noakes relied upon the judgment of Rix LJ in Scott v Copenhagen Re [2003] EWCA Civ 688; [2003] 2 All ER (Comm) 190 at [67]-[68]:

“67..  I revert to Mr Boyd's four principal submissions. The first (para 26 above) was that the causal link expressed by the phrase “arising out of” is a weak one, and that therefore and possibly in any event there was nothing that occurred after the initial invasion and capture of the airport that outweighed their significance for purposes of causation.

68.

In my judgment, however, there is nothing in the authorities to support that submission as a matter of principle. On the contrary, it seems to me that in Dawson's Field, and again in Caudle v. Sharp a significant causal relationship was, albeit implicitly, imposed. In those cases Mr Kerr and this court found the relevant event (or occurrence) in the nearer events, rather than in the more distant. In the latter case, the concept of remoteness was expressly adverted to. I accept that in Dawson's Field the choice was obscured by the fact that the hijackings could not be regarded as a single event, and for that reason could not even be a candidate. Nevertheless, it seems to me ultimately to be inherent in the concept of aggregation (“arising out of one event”) that a significant causal link is required. In this connection I would refer to Lord Hoffmann's substantial contribution in recent years to an understanding of what lies behind the courts' intuitive judgments on issues of causation: see, for instance, Empress Car Co (Abertillery) Ltd v. National River Authority [1999] 2 AC 22  at 29/35. Lord Hoffmann emphasises that it is not possible to give an informed answer to a question of causation when attributing responsibility under some rule without knowing the purpose and scope of the rule. In the present context, the purpose and scope of the rule has to be found in the concept of aggregation inherent in wording such as “arising out of one event”. A plurality of losses is to be regarded as a single aggregated loss if they can be sufficiently linked to a single unifying event by being causally connected with it. The aggregating function of such a clause is antagonistic to a weak or loose causal relationship between losses and the required unifying single event. This is the more easily seen by acknowledging that, once a merely weak causal connection is required, there is in principle no limit to the theoretical possibility of tracing back to the causes of causes. The question therefore in my judgment becomes: Is there one event which should be regarded as the cause of these losses so as to make it appropriate to regard these losses as constituting for the purposes of aggregation under this policy one loss?”

Mr Christie KC and Ms Noakes relied upon [69] of The FCA Test Case where in the context of claims under direct insurance policies, the majority of the Supreme Court held that neither a particular disease or an "outbreak of disease" constituted an occurrence:

“A disease that spreads is not something that occurs at a particular time and place and in a particular way: it occurs at a multiplicity of different times and places and may occur in different ways involving differing symptoms of greater or less severity. Nor for that matter could an 'outbreak' of disease be regarded as one occurrence, unless the individual cases of disease described as an "outbreak" have a sufficient degree of unity in relation to time, locality and cause. If several members of a household were all infected with Covid-19 when a carrier of the disease visited their home on a particular day, that might arguably be described as one occurrence. But the same could not be said of the contraction of the disease by different individuals on different days in different towns and from different sources. Still less could it be said that all the cases of Covid-19 in England (or in the United Kingdom or throughout the world) which had arisen by any given date in March 2020 constituted one occurrence. On any reasonable or realistic view, those cases comprised thousands of separate occurrences of Covid-19. Some of those occurrences of the disease may have been within a radius of 25 miles of the insured premises whereas others undoubtedly will not have been. The interpretation which makes best sense of the clause, in our view, is to regard each case of illness sustained by an individual as a separate occurrence. On this basis there is no difficulty in principle and unlikely in most instances to be difficulty in practice in determining whether a particular occurrence was within or outside the specified geographical area.”

Mr Christie KC and Ms Noakes submitted that, if the judge had put together that paragraph (which he cited in the context of the Markel appeal at [107] but did not refer to in the context of the Covéa appeal) with what he concluded in [104], he would have been bound to conclude that the outbreak of Covid-19 was not an event and therefore it and the pandemic could not be a “catastrophe”. The judge had committed a fundamental error of law in relation to what constituted a “catastrophe” by failing to have regard to the Supreme Court decision.

A number of the other contentions advanced by Unipol on the first ground of appeal were set out in its Skeleton Argument but not developed in oral submissions. I will summarise those contentions in the following paragraphs.

Unipol argues that a catastrophe must be a sudden and violent event. It submitted that the words used in each of the OED and SOED definitions connote suddenness, either expressly or impliedly. They also all expressly or impliedly define “catastrophe” as a type of event.

The Judge’s conclusions that the academic and market commentaries show that the word “catastrophe” has any meaning other than what it would bear on ordinary principles of construction do not bear analysis. Unipol submitted that while, in contrast with the LIRMA wording, the Covéa Reinsurance does not refer to an “immediate result” or to “sudden violent physical operation”, it requires losses to be “directly occasioned” by catastrophes and there was no conceptual difference between these various phrases.

Changes in the business written in the books of direct insurers are nothing to the point since this could have no bearing on the meaning of a Property Catastrophe Excess of Loss Reinsurance. Strong support for Unipol’s case that the focus in terms of “suddenness” should be on the emergence of the “catastrophe” was to be found in the judgment of the Full Federal Court of Australia in the Swiss Re case. Unipol also submitted that, contrary to the judge’s assertion that riots or civil commotion and floods need not be sudden, it is entirely natural to talk about a sudden outbreak of rioting or about there being sudden flooding when a levee breaks. Both events are sudden in that they are a “happening” at an identifiable time. If the judge had correctly concluded that a “catastrophe” required “suddenness” he could not have concluded that the outbreak as found by the tribunal constituted a “catastrophe”. The outbreak and spread of Covid-19 in the UK and the first lockdown were seen from some way out and surprised no one; the “explosion” of cases necessarily “emerged relatively slowly or progressively over time” (Swiss Re); it cannot be said to have happened “suddenly”.

Unipol submitted that the judge was wrong to reject each of its five arguments that the “catastrophe” must be one which causes, or is capable of causing, physical damage. It submitted that the extracts from Butler & Merkin and Kiln set out in the judgment provided strong support for its case. The judge’s reasoning for rejecting this at [74]–[75] (i.e. the evolution of non-damage BI extensions in direct policies) is fallacious. At [76], the judge said that, because claims for BI consequential upon loss of access to the insured property involved an interference with the original insured premises, they were “apt for inclusion”. Unipol submitted that this was nothing to the point absent express exclusion.

In relation to Condition 2(vii) of the Reinsurance viz. “any Loss Occurrence of whatsoever nature” at [78] the judge said that this was clearly intended to extend beyond unidentified members of the same nature as those specifically mentioned. Unipol submitted that there is no basis not to apply the ejusdem generis principle: a hurricane, flood, or riot are all physical events with physical effects, and the clause would still have meaning without opening the floodgates. Rarely would a drafter expressly define a genus. If they did, there would be no need for the ejusdem generis principle at all.

At [82], the judge suggested that Condition 18 was a specialist extension and provided no basis for reading down the remainder of the contract, and that a “client whose broker had “obtained” such an extension would be disappointed to learn that the remainder of their cover was narrower than if they had not had it”. Unipol submitted that this analysis assumes the very thing in issue. It is just as correct to say that the client who obtained such an extension would be pleased to have it, since it was not otherwise covered. If Unipol is correct that cover must be expressly provided by the Reinsurance, then Condition 18 strongly supports its case.

At [84]-[85], the judge said that he did not find the Transmission Exclusion “particularly informative”, partly “because it contemplates the operation of a peril which has caused and is capable of causing physical damage, albeit to the property of a third-party”. However, Unipol submitted that this reinforces the requirement for physical damage and supports Unipol’s case.

In relation to Ground 2 concerning the Hours Clause, Mr Christie KC and Ms Noakes submitted that, on the proper construction of Conditions 1 and 2, there were three steps for recovery under the Covéa Reinsurance: (1) Covéa must identify a single catastrophe; (2) it must demonstrate that a class of individual losses was directly caused by this catastrophe, thus identifying a class of losses that may aggregate and (3) it must show that the aggregate of individual losses occurring within the relevant period of time specified for the relevant catastrophe exceed the deductible, thus identifying the sub-set of losses that will aggregate. They submitted that step (3) follows steps (1) and (2) and is entirely mechanical. In no circumstances would individual losses that occurred outside of any Loss Occurrence period aggregate with losses occurring within that period; the final sentence of Condition 2(2)(i) eliminates any vestige of doubt: “… and no individual loss from whatever [catastrophe] which occurs outside [the relevant period] shall be included in that “Loss Occurrence”.

Mr Christie KC and Ms Noakes submitted that the Covéa Reinsurance thus provides an indemnity for individual losses that occur in carefully prescribed periods of time that were each directly caused by one and the same catastrophe. The suggestion that an individual loss on a particular day could somehow embrace other losses, which occurred subsequently, defeats the whole purpose of the Hours Clause for two reasons:

The fundamental error committed by the judge was that he failed to have proper regard to the terms of the Covéa Reinsurance. Had he done so he would have been bound to conclude that the “individual losses” were the losses sustained by any Covéa policyholders on a day-to-day basis, and that the function of the Hours Clause was to aggregate those losses and allocate them to a 168 hour “Loss Occurrence” under Condition 2(2)(vii).

Mr Christie KC and Ms Noakes submitted that the judge had been wrong in [148] (set out at [56] above) to conclude that, by analogy with the treatment of damage BI, an “individual loss” occurs under a non-damage policy when the insured peril strikes the premises, the peril being the loss of ability to use the premises (whether through damage to other property or a closure order). None of the judge’s five factors withstands analysis.

As to (i) they submitted that it is precisely the function of the Hours Clause to allocate loss to different “Loss Occurrences” but that does not involve arbitrary apportionment by reference to a fiction as to when a loss occurs which is the judge’s approach. As to (ii) whether or not there are differences between treatment of physical damage and non-damage BI losses at the direct level is nothing to the point. The case concerns aggregation of “individual losses” under the reinsurance which provides that “individual losses” are losses which occur at a particular time and place. The cases referred to in (iii) all concern direct policies with different aggregation wording. Further, the passages cited from Stonegate and Various Eateries are explicable based on the terms of the direct policies; neither case is concerned with whether a BI loss occurs day-by-day, or occurs on day 1; and neither is concerned with the treatment of loss under a Catastrophe XL Reinsurance. The passages cited from The FCA Test Case do not support the judge’s conclusion. If the “catastrophe” is the outbreak of Covid-19 in the UK leading up to 18 March 2020, the policyholders’ losses were not “directly occasioned” by that “catastrophe”, but by the closure of the insured premises.

They submitted that contrary to what the judge said in (v), Unipol’s “day by day” construction does not lead to “uncommercial consequences”. The matters relied upon by the judge at [138] to [142] are irrelevant to the construction of the Reinsurance. The way in which BI loss was calculated under direct policies says nothing about the meaning and operation of the Hours Clause. The interests of the direct insurers and reinsurers are not aligned; they have entered into completely separate bargains. The Hours Clause is there to delineate and limit reinsurers’ exposure.

The fact that direct insurers may calculate losses retrospectively taking into account variations during the indemnity period does not alter the fact that the losses occur day-by-day. If a direct insurer can calculate losses over a three-month period, it can do so over a period of one week or one hour without any conceptual difficulty. It may well be that the direct insured suffers no loss in the 168-hour period after the peril strikes and, in that case, there will be no recoverable loss under the Reinsurance.

Mr Christie KC and Ms Noakes submitted that the “practical difficulties” with the construction adopted by the tribunal and the judge which the judge dismissed at [150] (set out at [57] above) were in fact substantive legal issues which should have precluded him from reaching the conclusions he did. They cited the example of the business proprietors who were on vacation returning to the premises many months after the catastrophe struck and only then sustain business interruption because they are unable to operate the business on their return. The example negates the proposition that the “individual loss” occurs when the peril (i.e. the closure order) affects the premises. On any commonsense view and as a matter of law, no recoverable loss would have been sustained until their return.

The judge sought to dismiss that conclusion, saying at [150(ii)]: “the interference with the owner’s right to use the [insured premises] occurs when the order comes into effect, whatever use the owner wishes to make of the [insured premises] at that point in time.” Mr Christie KC and Ms Noakes submitted that this misses the point. There could be interference with a right to use property as soon as a restriction order comes into effect, but it may not give rise to loss, without which there can be no indemnity claim. It would be absurd to suggest that the insured could recover damages where the only loss was a hypothetical loss of use. On the contrary, a loss of use may result in a cost saving. A cost saving in, say, week 1, would preclude an indemnity for that week (to the extent the saving outstripped loss) because there would be no recoverable loss. Conversely, if savings in week 1 meant that there was no loss in week 1, but correspondingly greater losses in week 2, then those losses would be recoverable in week 2 (subject to the Hours Clause).

They submitted that this outcome is not surprising. It is precisely what Condition 2(4)(4) referred to by the judge at [150(vi)] mandates. It is not a question, as the judge thought of ““slicing and dicing” what, at the direct insurance level, is a net loss”; it is a question of applying the terms of the Reinsurance to the “individual losses” in chronological order. The Judge’s approach seeks impermissibly to impose on the Reinsurance terms in the direct policy relating to the calculation of loss.

They submitted that commercial considerations supported Unipol’s approach. On the Judge’s interpretation, (1) many parts of Condition 2 would be redundant, (2) the Hours Clause would be denuded of effect in relation to BI, and (3) Condition 3 would be overridden. The cover would effectively be converted into aggregate or stop loss cover, for which Covéa did not bargain or pay. Further, without the clarity provided by the Hours Clause catastrophe XL cover will be all the more difficult to rate. If Covéa is placed in difficulty regarding any discontinuity of cover, the remedy was in its own hands; the wording should still be given its common sense meaning, applying Arnold v Britton.

On behalf of Covéa, Mr Schaff KC and Mr Kerr in their written submissions submitted that the judge did not make the two fundamental errors alleged by Unipol. In relation to the first, they submitted that the judge did not conflate Property Catastrophe XL Reinsurance with stop loss/aggregate XL Reinsurance. He addressed the distinctions between the types of cover at [103] when addressing Unipol’s argument that, in contrast with other forms of reinsurance, it is not sufficient for the reinsured’s losses to be ‘catastrophic’, but necessary to identify a ‘catastrophe properly so-called, which is an appropriate basis for aggregating losses’. The judgment set out certain characteristics of a catastrophe in the relevant commercial and contractual context at [104] and held that, on the basis of the findings in the Award, it could not be said that those characteristics had not been satisfied.

They noted that the tribunal did not consider that it would be sufficient merely for a phenomenon to have a ‘catastrophic’ result on an insured’s business. The tribunal explicitly considered the nature, suddenness of onset, scale and ‘coherence’ (in development and effect) of the phenomenon as well as its ramifications, and found that the exponential increase in infections during the first three weeks of March 2020 amounted to ‘a disaster of sudden onset such as to qualify as a catastrophe’ and that it was a ‘catastrophe that affect[ed] commercial property…’. As the judge held at [106(ii)]): ‘the “outbreak”… can be fairly regarded as a coherent and discrete happening, with an existence, identity and “catastrophic character”….’

Covéa submitted that the second “fundamental error” alleged by Unipol, that the judge had assumed that the terms of the direct policies could be read across to the Covéa Reinsurance, was also entirely misconceived. The judge had been alive to the point that the reinsurance was not “back to back” with the insurances. This is clear from [122] to [126] of the judgment (albeit dealing with arguments by Gen Re in the Markel Reinsurance appeal about the Hours Clause in that reinsurance) where the judge cited passages from the judgments of Butcher J in Stonegate and Various Eateries Trading Ltd v Allianz Insurance plc [2022] EWHC 2549 (Comm) and the Court of Appeal in the latter case ([2024] EWCA Civ 10; [2024] Bus LR 810) to the effect that the forced closure was a trigger with continuous effect rather than involving a series of day-by-day triggers. At [126] the judge then said:

“Relying on this analysis does not involve interpreting the aggregation provisions in the Reinsurances on the flawed premise that they are intended to operate in the same manner as those in the direct insurance (cf Axa Reinsurance (UK) Ltd v Field [1996] 1 WLR 1026, 1034). The reference to "individual losses" in both Reinsurances naturally directs attention to the position at the level of the original insured, and the Markel tribunal's overriding consideration – that "it is natural to think that business interruption losses occur day by day" (Markel Award, [58]) – is also concerned with the nature of such losses at the position of the original insured.”

They submitted that, far from committing a fundamental error, in an entirely orthodox and necessary appraisal of the relevant contractual context, in accordance with the principles of construction the judge set out [36] to [41] (which are not challenged on appeal), the judge considered the general nature of the risks typically written in a reinsured UK property account at the time that the Covéa Reinsurance was concluded at the end of 2019 and the typical operation of the underlying reinsured business interruption cover. This exercise was all the more apt where, as was common ground, the reference to ‘individual losses’ in Condition 2(1) of the Reinsurance was to the loss sustained by the original insured, which as the judge said at [147]: “points the enquiry in the direction of the direct insurance”.

Mr Schaff KC and Mr Kerr submitted that by contrast, Unipol’s approach to construction is heterodox and would entail the perverse consequence that uncontroversial contemporaneous market context would be excluded from consideration and the reinsurance would be construed exclusively with regard to historic materials. There is no ‘failure of logic’ in having regard to the market context in 2019, nor does this involve enlarging the proper meaning of the term “catastrophe”. This was a point the judge made at [58] (quoted at [22] above).

They also submitted that Unipol fundamentally misinterprets the analysis and decision in Axa v Field. Covéa has never suggested that the construction of aggregation provisions is to be ‘back to back’ or as having the same effect (the vice identified by Lord Mustill in that case). On the contrary, at the inwards level, aggregation is likely to have been on a ‘per occurrence’ basis, as in Stonegate, Various Eateries and Greggs, which is conspicuously not the case in this Reinsurance. Unipol appears to treat Axa v Field as if it were an authority for the bold proposition that the terms or subject-matter of the direct policy cannot even be used ‘to inform the interpretation’ of the reinsurance, as a matter of relevant contractual context. Axa v Field is no authority for such an extraordinary submission.

Mr Schaff KC and Mr Kerr submitted in relation to the first ground of appeal that the starting point was that “catastrophe” is not a defined term in the Covéa Reinsurance and has no recognised meaning in insurance law as the tribunal correctly found. They noted that it was expressly agreed between the experts who gave evidence in the arbitration that ‘there is not a common market-wide understanding or definition of what constitutes a catastrophe’.

As to the dictionary definitions, they submitted that, as a matter of ordinary language, any reasonable person would characterise what took place in the UK in the first half of March 2020 as ‘[a] sudden or widespread or noteworthy disaster, an extreme misfortune’ (SOED definition,),or ‘[a] sudden disaster, wide-spread, very fatal, or signal’ (OED definition 4),or‘[a]n event producing a subversion of the order or system of things’ (OED definition 3). Applying those definitions, the tribunal’s evaluative conclusion that the outbreak of Covid-19 in the UK, reflected by the exponential increase in the number of infections during the first three weeks of March 2020, amounted to ‘a large-scale national disaster of sudden onset that may accurately be described as a catastrophe’ is plainly correct.

Mr Schaff KC and Mr Kerr noted that whilst dictionary definitions may be helpful, the tribunal correctly recognised that “context is crucial to meaning”. They relied upon a number of points of context. First, the findings of the tribunal on the basis of agreed or unchallenged expert evidence on market practice and understanding concerning the writing of non-property damage BI risks, findings which were adopted by the judge at [60] (set out at [23] above). They also noted that it was undisputed between the experts that common types of cover for such risks included cover for BI losses caused by Notifiable Disease and/or by Prevention of Access to premises as the result of government action or advice following an emergency.

Second, they relied upon the fact that the Reinsurance was described as “Property Catastrophe Excess of Loss Reinsurance”, pointing out that the tribunal had correctly held and the judge agreed that this description ‘is consistent with the provision of cover… for a catastrophe capable of affecting property by preventing access to it and thereby causing its loss of use’ (emphasis added).

Third, the Reinsurance expressly provided in the “Class” provision: “This Contract shall indemnify the Reinsured in respect of all business written within the Reinsured’s Property Department and classified as … Commercial”. They submitted that the phrase “all business” meant what it said and encompasses and would have been understood as encompassing non-damage BI risks including Notifiable Disease risks and Prevention of Access risks (such as those insured by clause 7 of the NurseryCare policy). The judge rightly regarded the contractual provisions which he set out at [61] (cited at [24] above) as important and as providing ‘strong support’ for rejecting the supposed limitation. In his oral submissions Mr Schaff KC emphasised that there was no definitional requirement in the reinsurance for a “catastrophe” to mean only property damage.

Fourth, they submitted that, in using the term ‘catastrophe’ and not ‘event’ or ‘occurrence’, the Loss Occurrence clause adopted what the tribunal rightly considered to be ‘a more flexible concept [with] a wider compass’. This is hardly surprising when dealing with the aggregation of a multitude of original ‘occurrence’ based losses from business written in the Reinsured’s Property Department to recover under a Property Catastrophe XL Reinsurance providing cover (in three layers) for £140 million excess of £10 million. In his oral submissions, Mr Schaff KC made the point that, as the tribunal found, the market was alive to the distinction between “event” and “catastrophe”.

Fifth, as the arbitrators rightly observed, ‘it is explicit that the particular catastrophes identified in Condition 2… do not observe the unities applicable to an event or occurrence as laid down in the case law.’ Mr Schaff KC and Mr Kerr submitted that the tribunal’s analysis of typical loss occurrence examples (hurricanes and Australian bushfires), and of the way in which the Hours Clause in the reinsurance deals with those examples (and with riots or burst pipes/snow) as supporting their view, is impeccable.

In his oral submissions, Mr Schaff KC emphasised that an appeal under section 69 of the Arbitration Act 1996 was only concerned with whether there were errors of law. This was a point which was recently emphasised in this Court by Males LJ in Various Eateries at [59], citing the judgment of Sir Jeremy Cooke in Simmonds v Gammell:

“Some further support for this view in the present context can be derived from Simmonds v Gammell, where Sir Jeremy Cooke was right to consider that the relevant question was whether the arbitrators were "entitled" to decide the case as they did. Although an appeal from arbitrators involves the additional feature that the court's jurisdiction is limited to determining a question of law arising out of the award, and that it has no jurisdiction to review the arbitrators' factual findings, Sir Jeremy Cooke was right in my view to regard the issue as requiring the kind of exercise of judgment with which an appellate court will not interfere in the absence of an error of principle.”

In relation to Unipol’s criticism of the judge that a “catastrophe” must be an event or species of event, Mr Schaff KC and Mr Kerr submitted that the true gravamen of this complaint was not that the tribunal and the judge committed an error of law but that they respectively reached and confirmed an evaluative judgment on the facts that the sudden outbreak of Covid-19 in the UK was a catastrophe. They submitted that it was not open to Unipol to say that the outbreak was a “state of affairs”. The tribunal having made an evaluation of the facts at [49] of the Award as to the nature of the outbreak, it was not open to Unipol to challenge that on appeal:

“if the idea of a ‘sudden disaster’ is inherent in the meaning of catastrophe…the exponential increase in Covid-19 infections in the UK during the first three weeks of March 2020 did amount to a disaster of sudden onset such as to qualify as a catastrophe.”

Mr Schaff KC and Mr Kerr noted that the Covéa Reinsurance does not use the word “event” at all nor “occurrence” save in the phrase “Loss Occurrence” which is itself defined by reference not to “event” or “occurrence” but “catastrophe”. There is no basis for the suggestion that the flexible, undefined concept of “catastrophe” should be treated as a species of “occurrence” or “event” with all the accumulated legal baggage those expressions carry. In any case, even if a catastrophe may be properly described as a particular type of event, as the tribunal held: ‘it still requires a more flexible construction and a wider exercise of judgment to be applied than that which the jurisprudence stipulates for the word “event” simpliciter’. The judge had put it at [102] that the terms of the Reinsurance ‘point strongly away from anything but the most generous application of two of Lord Mustill’s three unities’, i.e. time and place, as the Hours Clause indicates that a qualifying catastrophe may be of potentially very broad temporal scope and wide geographical impact.

They also submitted that Unipol’s attempt to characterise the outbreak as a “state of affairs” or a description of or judgment on a state of affairs was simply inconsistent with the tribunal’s findings of fact and evaluation of those facts, specifically the finding quoted at [102] above. Contrary to Unipol’s submissions (summarised at [65] above) the tribunal held expressly that: “during this relatively short period, the COVID-19 outbreak assumed a certain coherence in its development and effect and gave rise to a profound subversion of the order of life within the UK” (emphasis added). This evaluation does not remotely accord with Unipol’s complaint that the catastrophe identified was “hopelessly vague”.

Mr Schaff KC and Mr Kerr submitted that Unipol’s related contention that a catastrophe must be a “sudden and violent event” should be rejected for the reasons given by the judge at [89] to [97] (summarised at [33] to [39] above). Any such requirement would be inconsistent with the Hours Clause under which many of the perils identified need not be sudden in their inception, short in their duration or violent in their impact or operation. In any event, any definitional requirement of “suddenness” would not advance Unipol’s case because as the tribunal correctly held: ‘in this context the word “sudden” has to be interpreted with relative flexibility’ and ‘On such an approach it is evident that the exponential increase in Covid-19 infections in the UK during the first three weeks of March 2020 did amount to a disaster of sudden onset such as to qualify as a catastrophe’. As the judge held at [97]: “There was no suggestion that this conclusion was ‘necessarily inconsistent’ with the correct application of the relevant legal test, nor could that submission have realistically been advanced”.

They submitted that Unipol’s former primary argument that the “catastrophe” must cause or be capable of causing physical damage and that this was ‘inherent’ in the meaning of “catastrophe” was correctly rejected by the judge. It was common ground that the ordinary meaning of the word ‘catastrophe’ contained no such inherence and, on the basis that it was agreed between the experts that there was no common market-wide understanding or definition as to what constitutes a catastrophe, it followed that there was no such common market-wide understanding or definition to support such inherence. They submitted that the tribunal rightly held that the background materials relied upon by Unipol did not advance its case. Mr Kiln himself recognised that: ‘it is impossible to envisage all the forms future catastrophes will take’. The tribunal found correctly that the content and scope of the typical property book have changed since the 1960s, and that ‘the nature of a catastrophe capable of causing multiple loss’ could also in principle change. By 2019, Mr Kiln’s historic views of the market and the background to the drafting of the clause were well out of date.

Mr Schaff KC and Mr Kerr submitted that Unipol’s complaint that non-damage BI losses arising from denial of access required ‘express inclusion’, puts the matter back to front. In light of the wide, unrestricted subject-matter of the Reinsurance (‘all business written within the Reinsured’s Property Department’), it was incumbent on Unipol expressly to exclude such losses if it wished to do so.

Unipol also sought to rely on an ejusdem generis argument. Mr Schaff KC and Mr Kerr submitted that thisis a tool of construction which may or may not apply depending on context. It was rightly rejected by both the Covéa and Markel tribunals, with which the Judge expressed full agreement. They highlighted the very wide and unrestricted words of Condition 2(2)(vii) in the Hours Clause: ‘any Loss Occurrence of whatsoever nature which does not include individual loss or losses from any of the insured perils [previously enumerated]’. As the tribunal found, the Hours Clause ‘is not attempting to create a class or genus but is simply ascribing hours to specific, well known catastrophes’ with that concluding provision in (vii) having a deliberately wide ambit to cover other non-enumerated catastrophes. The inclusion of catastrophes that prevent access is also supported by commercial sense, when viewed against the risks now typically covered by a property insurer, the clause’s ‘deliberately wide ambit’, and a reinsurance which includes non-damage BI.

They also submitted that Unipol’s suggested construction would be entirely arbitrary. It appears to be accepted that the peril only need be capable of causing physical damage. There is no rational basis for restricting the very wide language to cater only for those perils capable of causing physical damage to property, as opposed to impairing the physical use of or access to property to save lives.

As for Unipol’s reliance on Condition 18 (‘Destruction by Civil Authority’), they submitted that the clause concerns the scenario where loss or damage is deliberately inflicted upon insured property (pursuant to civil authority order) to protect against a peril. It has nothing to do with non-damage BI loss and provides no basis for ‘reading down’ the scope of the Reinsurance.

In relation to the second ground of appeal concerning the Hours Clause Mr Schaff KC and Mr Kerr submitted that the Covéa tribunal’s interpretation of that clause was entirely correct as a matter of construction, commercial sense and long-standing market practice and understanding concerning the operation of the clause in relation to damage BI losses. As the judge found, the contrary decision of the Markel tribunal was wrong.

They pointed out that the general operation and effect of the Hours Clause was either common ground or undisputed in the arbitration. The tribunal found that: (i) its purpose is to limit the duration and extent of a Loss Occurrence by confining the aggregation of ‘individual losses arising out of and directly occasioned by one catastrophe’ to those ‘individual losses’ which occur within a specified period applicable to the peril in question (e.g. earthquake/riots: 72 hours; hurricane: 120 hours; flood: 504 hours; other unspecified perils: 168 hours); (ii) ‘Individual losses’ must be allocated to any Loss Occurrence according to when any such individual losses occur. Those which occur outside the specified/selected period of hours form part of a separate Loss Occurrence subject to a separate deductible; (iii) The ‘individual loss’ which must fall within the specified Loss Occurrence window is the loss under the original insurance.

In his oral submissions, Mr Schaff KC emphasised what the tribunal and the judge had found, that the Hours Clause limits the duration and extent of the Loss Occurrence as defined, not the duration and extent of any individual loss that occurs within a Loss Occurrence period. What the Hours Clause requires is that the individual loss occurs within the Loss Occurrence period (here 168 hours) but if it continues to be sustained thereafter, there is nothing in the clause which requires the individual loss to be apportioned so that only the part of it sustained during the 168 hours is to be indemnified.

Mr Schaff KC and Mr Kerr submitted that the material question in every case of aggregation under the Hours Clause is: when did the particular (original) ‘individual loss’ occur? A date of occurrence has to be ascribed to each ‘individual loss’ in order to resolve whether it occurred inside or outside the specified period of hours. “Occur” in context means when did any individual loss first occur, even if it continued to develop over time. They submitted that this was consistent with the terms of the reinsurance: not just the Hours Clause itself, but the Reinstatement and Extended Expiration provisions to which the judge referred at [144], both of which require a specific date to be ascribed to an individual loss.

They submitted that it is also consistent with the first principles in “losses occurring during” (re)insurance cover where the loss is attributable to the policy year in which it first occurs. It is also consistent with the Covid-19 jurisprudence: in Various Eateries both Butcher J (at [67] and [69]) and the Court of Appeal (at [93-[95]) held that provided there was an enforced closure of premises within the period of cover, the insured could recover the full amount of its proximately caused losses even if the resulting closure extends beyond the policy period.

So far as cases of damage BI losses are concerned, it was undisputed between the experts that ‘market practice was and is to treat the [BIL] as occurring simultaneously with the material damage’. The critical questions of construction are: (i) whether that market practice is correct as a matter of principle; and (ii) whether the same approach should apply in cases of non-damage business interruption loss. The tribunal answered both of those questions in the affirmative and the judge held at [148]-[151] that they were entirely right to do so.

Mr Schaff KC and Mr Kerr submitted that both were correct. It was common ground that the phrase “individual loss” meant loss sustained by the original insured. They submitted that the meaning of the phrase does not change depending upon whether the relevant “individual loss” comprises physical damage losses, BI losses or both. The phrase encompasses the entirety of the loss sustained by the original insured as a result of the relevant catastrophe striking the insured property, without differentiation. As the judge concluded, treating the insured’s loss as a single ‘individual loss’ for the purposes of reinsurance recovery (rather than ‘slicing and dicing’ it into segments of 168 hours or less) reflects the treatment of that loss at direct level.

They submitted that the central question for the application of the Hours Clause is the same for BI losses (whether damage or non-damage) as for property damage losses: when does the relevant “individual loss” occur? To be aggregated it must occur, that is, first occur, within the relevant period of hours. The focus of the clause on when ‘individual losses’ occur and whether they occur inside or outside the specified period of hours, mandates this approach to BI losses. The fact that one is dealing with BI losses does not alter the contractual position: there is still no basis for apportionment of an “individual loss” or its financial elements into parts suffered within the period of hours and parts suffered outside the period of hours.

They submitted that, although there is a factual difference between an ‘individual (damage BI) loss’ and an ‘individual (non-damage BI) loss’ (deriving from the different type of impact of the catastrophe), there is no analytical difference for the purposes of the timing of an occurrence of an ‘individual loss.’ The tribunal rightly held that an ‘individual loss’ is ‘a loss sustained by an original insured which occurs as and when a covered peril strikes or affects insured premises or property’, it being immaterial for these purposes how the property is affected and by what type of peril. In all such cases, an ‘individual loss’ still only occurs once for the purposes of the Hours Clause, notwithstanding any subsequent continuation.

This approach to timing accords with the operation of the Reinsurance generally. A single date of an ‘individual loss’ needs to be ascribed for coverage and reinstatement purposes. They submitted that, by contrast, Unipol’s case that BI losses paid by the reinsured ‘have to be unpicked at the reinsurance level’ is, as the judge found, likely to produce ‘particularly acute’ difficulties, involving the ‘slicing and dicing’ of what, at the direct level, will have been a single net loss. As the judge observed at [142], the assessment of the amount paid to settle an individual BI loss entails looking at the net effect of credits and debits over time, something ‘very far-removed from a “day to day” calculation’.

Mr Schaff KC and Mr Kerr submitted that, viewed more broadly, on Unipol’s case, a catastrophe reinsurance programme operating excess of £10m each and every Loss Occurrence in respect of ‘all business’ written in Covéa’s property department would hardly ever respond to any BI losses (damage or non-damage) if all such BI losses have to be divided into such small periods of time, leaving Covéa unreinsured for original BI losses subject to maximum indemnity periods of three months or longer, and producing an uncommercial decoupling of the insurance and reinsurance protections for the same underlying ‘individual losses’.

They submitted that the judge’s reasons for agreeing with the analysis and conclusions of the Covéa tribunal were unassailable. Unipol’s central submissions are that BI losses ‘accrue day by day’; that losses ‘occur at specified times’; and, that an individual loss cannot extend ‘beyond the relevant period of hours’; and that it would be ‘flatly inconsistent’ with the meaning and intention of the clauseto conclude otherwise. These submissions are no more than assertions, unsupported by and inconsistent with the language, the Reinsurance more generally, and the wider commercial context.

More specifically, Unipol confuses the occurrence of an ‘individual (BI) loss’ with the continuing financial loss the subject of such individual loss. The interference with the business of each nursery school which occurred on closure on 20 March 2020 is functionally equivalent to each school suffering property damage on that day. Whether or not one could predict at the outset the time period the premises would be closed and the amount of resulting loss, the relevant individual loss occurred on day 1 and not day-by-day for every day that the business interruption continued or the continuing financial loss was sustained.

This analysis does not, as Unipol suggests, ‘generate a cause of action where there is no loss’ or undermine the ‘fundamental principle of indemnity’. The closure order amounted to an immediate interference. Even if it might be said that factually a business with vacationing proprietors might not suffer an immediate interference and only suffer an individual loss on its scheduled re-opening, it would still only sustain a single individual loss on that later date, rather than on a day by day basis; and, if an original insured never suffered a financial loss at all, then its claim could never fall for inwards recovery, let alone outwards aggregation. In any event, these hypothetical scenarios are moot so far as the insured nurseries are concerned. They did each suffer an individual loss when they were first ordered to close and did close on 20 March 2020.

Finally, Mr Schaff KC and Mr Kerr addressed what they described as the “miscellany of alarmist points” advanced by Unipol, referred to in summary at [87] above, none of which had any merit.

It is appropriate to consider first the two “fundamental errors” which Unipol alleged the judge made in his analysis of the issues, both of which I consider to be unmeritorious. The submission that as experienced a Commercial Court judge as Foxton J had conflated the nature of cover under a Property Catastrophe XL Reinsurance with cover under a stop Loss/aggregate XL reinsurance was never a promising one and so it proves. At [52] (cited at [17] above) the judge identified the three different types of reinsurance protection available, by reference to Mr Kiln’s book and there is no question of his having confused or conflated them. For good measure he returned to the distinction between property catastrophe XL cover and stop loss cover at [103].

In relation to the second alleged fundamental error, that the judge wrongly assumed that the terms of the underlying direct policies could be read across into the reinsurance, he did nothing of the sort. He was well aware that the Covéa Reinsurance and the underlying insurances were not back to back and should not be interpreted on the basis that they were intended to operate in the same manner, a point he made expressly at [126] quoted at [90] above. Rather, as Mr Schaff KC and Mr Kerr correctly submitted, in accordance with orthodox principles of construction summarised at [36] to [41] of the judgment, the judge considered the general nature of risks typically written in an insurer’s UK property account when the Covéa Reinsurance was concluded in late 2019 and the typical operation of BI cover under such insurances. To have ignored those matters and somehow considered the Covéa Reinsurance in isolation would have been misconceived. Contrary to what Unipol appears to have been contending, there is nothing in Axa v Field which dictates that approach. Furthermore, given that the reference to “individual losses” in Condition 2(1) of the Reinsurance was clearly to losses sustained by the original insureds, as the judge said at [147]: this “points the enquiry in the direction of the direct insurance.”

In my judgment the judge’s conclusion, in accordance with what the tribunal had found, that the outbreak of cases of Covid-19 in the UK in the period immediately preceding the closure of schools and nurseries on 20 March 2020 was a “catastrophe” was plainly correct.

The Covéa Reinsurance did not contain a definition of “catastrophe” and it was common ground that there is no special market definition or meaning, in contrast with “event” or “occurrence” both of which come in the context of reinsurance with what Mr Schaff KC and Mr Kerr described as “accumulated legal baggage”. Neither of those words is used in the Reinsurance other than in the phrase “Loss Occurrence” which is defined in Condition 2(1) as “all individual losses arising out of or directly caused by one catastrophe” so is referring back to “catastrophe”. In the circumstances, it is appropriate to have regard to the meaning of “catastrophe” in ordinary language albeit by reference to the contractual context.

The judge considered the dictionary definitions in the OED and SOED and summarised the position at [70] (set out at [26] above). He noted that both dictionaries offered meanings which did not require “suddenness” and that many of the definitions emphasise the existence of a significant break with the position up to that point and something which is seriously adverse in its nature or effects. Both those definitions seem to me apt to describe the outbreak of Covid-19 in the UK in March 2020.

The judge was also correct, as was the tribunal, to consider the dictionary definitions against the contractual and market context. The reinsurance wording was of particular significance. The Covéa Reinsurance was described as “Property Catastrophe Excess of Loss Reinsurance” which, as the tribunal and the judge considered: ‘is consistent with the provision of cover… for a catastrophe capable of affecting property by preventing access to it and thereby causing its loss of use’. The Class provision in the Reinsurance indemnified Covéa in respect of “all business written within [its] Property Department and classified as…Commercial.” Under Special Condition (ii) Covéa was to be the sole judge as to what was classified as “Commercial” business. As the tribunal and the judge found (see [2] above) the cover provided by the Covéa NurseryCare wording covered a wide range of risks commonly found in commercial insurance written in an insurer’s property department including cover for non-damage BI. Whatever may have been the position historically, BI cover written at the time that the Covéa Reinsurance was concluded in late 2019 commonly included cover for non-damage BI, such as losses caused by Notifiable Disease and/or Prevention of Access. This was the agreed or unchallenged evidence of the experts recorded by the Tribunal and adopted by the judge at [60(iii)]. The judge correctly regarded these contractual provisions as providing strong support for there being no limitation on the scope of a “catastrophe.”

Before this Court as before the judge, Unipol argued that there were three aspects of the meaning of the word “catastrophe” which it submitted the tribunal (and in turn the judge) had failed to recognise. The first of these was the point which became most significant in the oral submissions of Mr Christie KC, that a “catastrophe” must be an event or species of event which the outbreak of Covid-19 was not because it was a “state of affairs”. In my judgment, there are a number of problems with this argument. To begin with, as already noted at [129] above, the word “event” is not used anywhere in the Covéa Reinsurance and, if it had been intended that “catastrophe” was to be synonymous with “event” the Reinsurance would have said so or used the words “one event” not “one catastrophe” in the aggregating provision in Condition 2(1).

The word “occurrence” is also not used in the Covéa Reinsurance other than in the phrase “Loss Occurrence” which is defined in the aggregating provision in Condition 2(1) by reference to “one catastrophe”. In the circumstances, it is nothing to the point what “occurrence” means in underlying contracts of insurance such as those being considered by the Supreme Court at [69] of the majority judgment in the FCA Test Case cited at [67] above. The Supreme Court was not considering the meaning of “catastrophe” or of aggregating provisions in contracts of reinsurance.

The contention by Unipol that the outbreak of Covid-19 was a “state of affairs” is evidently drawing on the distinction drawn by Lord Mustill between “event” and “originating cause” in describing the three “unities” in Axa v Field at 1035:

“In ordinary speech, an event is something which happens at a particular time, at a particular place, in a particular way. I believe that this is how the Court of Appeal understood the word. A cause is to my mind something altogether less constricted. It can be a continuing state of affairs; it can be the absence of something happening.”

However, as already noted, this Reinsurance does not use the word “event” nor does it use “originating cause” and, in any event, as Sir Jeremey Cooke said in Simmonds v Gammell at [29]: “The "unities" are merely an aid in determining whether the circumstances of the losses involve such a degree of unity as to” satisfy the contractual aggregation requirement. In applying that principle, the judge correctly recognised at [102] that there were factors which pointed strongly away from anything other than the most generous application of two of the unities i.e. time and place. As set out at [42] above those factors included that Condition 2(2)(v) contemplated that a flood could have a duration of 504 hours (three weeks) and still be a catastrophe and that the Australian bushfires developed in a number of locations over weeks if not months and were still agreed to be a catastrophe. The judge went on to determine at [106(ii)]:

“In the context of the Covéa Reinsurance, the "outbreak" described by the Covéa tribunal can fairly be regarded as a coherent and discrete happening, with an existence, identity and "catastrophic character" which arise independently of the fact that it causes substantial losses. As the Covéa tribunal noted, "during this relatively short period, the Covid-19 outbreak assumed a certain coherence in its development and effect and gave rise to a profound subversion of the order of life within the UK" ([49]).”

I agree with Mr Schaff KC and Mr Kerr that this is an evaluative judgment by the judge that the sudden outbreak of Covid-19 in the UK in March 2020 was a “catastrophe” and that the gravamen of the complaint being made by Unipol is not that the tribunal or judge made an error of law but that they reached this evaluative judgment on the facts. The tribunal made a number of other important findings which the judge adopted including that: “the outbreak of Covid-19 in the UK reflected in an exponential increase in the number of infections during a period up to and including 18 March 2020” and “during this relatively short period, the COVID-19 outbreak assumed a certain coherence in its development and effect and gave rise to a profound subversion of the order of life within the UK”.

These findings also clearly involve an evaluative judgment on the facts as they were found by the tribunal and do not involve any error of law, from which it follows that they cannot be the subject of an appeal to this Court, as Males LJ said in Various Eateries at [59] in the passage cited at [100] above. The outbreak as described by the tribunal and the judge cannot, by any stretch of the imagination, be described as a “state of affairs”.

The second aspect of the meaning of “catastrophe” which Unipol submitted that the tribunal and the judge had failed to recognise was that a “catastrophe” must be a sudden and violent event. The judge noted, by reference to the dictionary meanings, that although some of them involved sudden happenings, they were not limited to such meanings. Furthermore, many of the perils identified in the Hours Clause need not be sudden in their inception or violent in their impact or operation. As the judge pointed out, riot or civil commotion, although violent, need not be sudden. They may develop over time and need not be short in duration. Floods may be sudden, as in flash floods or a dam bursting, but they can also develop over time, with gradually rising water and that sort of flood could hardly be described as violent.

In any event, even if there were a requirement of suddenness as part of the definition of a “catastrophe”, the tribunal held that it would have been satisfied. Its finding was: ‘in this context the word “sudden” has to be interpreted with relative flexibility’ and ‘On such an approach it is evident that the exponential increase in Covid19 infections in the UK during the first three weeks of March 2020 did amount to a disaster of sudden onset such as to qualify as a catastrophe’. I agree with Mr Schaff KC and Mr Kerr that the judge was correct in saying that this conclusion was not necessarily inconsistent with the application of the correct legal test.

The third aspect of the meaning of “catastrophe” which Unipol submitted that the tribunal and the judge had failed to recognise was that a “catastrophe” must cause or be capable of causing physical damage. This was Unipol’s primary case before the judge, but given the judge’s analysis of the breadth of the reinsurance wording, which I have summarised at [131] above, it is perhaps not surprising that this argument has been relegated to third place. It was common ground that this requirement was not inherent in the ordinary meaning of the word “catastrophe” and, given that it was agreed between the experts that there was no common market understanding or definition as to the meaning of “catastrophe”, there could not be any common market understanding or definition that this requirement was inherent in the concept of a “catastrophe”. Mr Schaff KC and Mr Kerr were also correct that the background materials Unipol had relied upon did not assist their case. Just as the content and scope of a typical property insurance account have changed since the 1960s when Mr Kiln first wrote his book, so the nature of a “catastrophe” may well have changed. Mr Kiln himself had expressly recognised this when he said: ‘it is impossible to envisage all the forms future catastrophes will take’.

The attempt by Unipol to rely on an ejusdem generis argument to support its case that a “catastrophe” must cause or be capable of causing physical damage is misconceived. Condition 2(2)(i) to (vi) in the Hours Clause are not purporting to set out a defined class, but ascribing hours to specific recognised catastrophes. The wording of Condition 2(2)(vii): “any Loss Occurrence of whatsoever nature which does not include individual loss or losses from any of the insured perils [in (i) to (vi)]” is extremely wide and intended to encompass other non-identified catastrophes. Nothing in the provision suggests that it is limited to perils of the same class or kind as (i) to (vi). Instead “of whatsoever nature” indicates it is intended to cover all other catastrophes, irrespective of whether they are of the same class or kind. The provision is clearly intended to capture risks now typically covered by a property insurer, to the extent they are not identified earlier in Condition 2(2).

I also agree with Mr Schaff KC and Mr Kerr that Unipol’s construction is entirely arbitrary. Since it seems that it is accepted, even on its construction, that the peril need only be capable of causing physical damage there is no rational ground for limiting the wide wording of (vii) to perils capable of causing physical damage and excluding perils which impair physical use of or access to property.

Unipol’s reliance on Condition 18 is misplaced. As the judge correctly found, it is a specialist extension to cover where loss and damage is deliberately inflicted on insured property pursuant to the order of a civil authority. It has nothing to do with non-damage BI loss and does not support any attempt to read down the wording of the Reinsurance.

In conclusion on the first ground of appeal, in my judgment, the judge’s analysis, culminating in his conclusion at [105]-[106] quoted at [45] above, as to why the outbreak of Covid-19 was a “catastrophe” within the meaning of the Covéa Reinsurance cannot be faulted.

In relation to the second ground of appeal, I agree with Mr Schaff KC and Mr Kerr that the central question for the application of the Hours Clause is when does the relevant “individual loss” (which it is common ground means the loss suffered by the original insured) occur, whether that is a property damage loss or a BI loss (damage or non-damage), since if the individual loss occurs outside the relevant period of hours, it cannot be included in the “Loss Occurrence”, as the closing words of the Hours Clause make clear. I also agree with their analysis that “occur” here must mean “first occur” so that what can be aggregated is individual losses which first occur during the relevant period, here the 168 hours or one week, even if the financial loss in question continues to develop over time after the 168 hours has expired.

In my judgment, the analysis of the judge at [148], which followed the analysis of the Covéa tribunal, is correct. 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 [148(ii)] I can see no basis for treating non-damage BI losses differently from damage BI losses. As he said: “there are obvious parallels between the impairment of the rights of those entitled to property resulting from damage and that resulting from the inability to use the property.”

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. There is nothing in the Hours Clause which requires the individual loss to be apportioned so that only the part of it sustained during the 168 hours is to be indemnified.

This analysis accords with the other provisions of the Reinsurance to which the judge referred at [144]-[145]. It also accords, as Mr Schaff KC and Mr Kerr submitted, with the first principles of “losses occurring during” (re)insurance cover where the loss is attributable to the policy year in which it first occurs. It is also consistent with the Covid-19 jurisprudence, albeit previous decisions of the Courts have concerned direct insurances rather than reinsurance. Specifically, this Court held in Various Eateries at [93] to [95] that, provided that there was an enforced closure of premises within the period of cover, the insured could recover the full amount of its losses even if the closure extended beyond the policy period.

By contrast, I consider that Unipol’s approach to the Hours Clause, which is that the individual losses occur day-by-day and that the Hours Clause limits the individual losses which can be aggregated to those which occur in the 168 hour period, so that only business interruption during that 168 hours is covered, is artificial and gives rise to considerable practical difficulties. As the judge said at [150(v)], profits costs and savings are likely to be “lumpy” in their effect rather than linear over time, so that the BI loss actually suffered may not become clear for some considerable time. As the judge went on at [150(vi)], Unipol’s approach would involve “slicing and dicing” in the Reinsurance what, at direct insurance level, is a net loss arrived at taking account of debits and credits over the entire indemnity period into account. As the judge said at [142] by reference to the passage in the majority judgment of the Supreme Court in the FCA Test Case which he had quoted at [141]:

“It will be apparent from the above that the amount paid to settle an individual business interruption loss can reflect a combination of credits and debits over a certain period, and that there may be considerable variation over time before you arrive at the final amount. This is very far-removed from a "day by day" calculation which UnipolRe's and General Reinsurance's arguments appear to assume. It is also clear that the assessment of a Business Interruption loss at the direct insurance level involves looking at the net effect over a particular period, not the aggregation of a series of daily losses…”

I agree with Mr Schaff KC and Mr Kerr that, on Unipol’s approach, this Property Catastrophe Excess of Loss Reinsurance, which provided cover in three layers, the lowest of which was £20 million in excess of £10 million Ultimate Net Loss each and every Loss Occurrence in respect of all business written in Covéa’s property department, would hardly ever respond to BI losses (whether damage or non-damage) since those BI losses would have to be divided into small periods of time, here no more than 168 hours. The existence of the reinstatement provision would not assist, since Covéa would be unable to aggregate individual losses getting anywhere near the lower limit of the reinsurance protection.

As Mr Schaff KC and Mr Kerr submitted, Unipol’s approach essentially conflates the occurrence of an “individual loss” with the continuing financial loss to which that individual loss gives rise. As they put it, 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.

In conclusion on the second ground of appeal, in my judgment, the judge’s analysis cannot be faulted and this ground of appeal must also be dismissed.

I agree.

I also agree.