Landmark Commercial Court decision on COVID-19 BI loss aggregation

An excess of loss reinsurance contract is one under which reinsurers will be liable to indemnify in the event that the reinsured has suffered a loss (typically defined as a "Loss Occurrence" ^[1] ) in excess of the stated attachment point. To this end, almost all such contracts will contain a Loss Occurrence clause (whether so labelled or by some other name) containing two elements. The first is concerned purely with aggregation, defining the Loss Occurrence as the cohort of all individual losses arising out of and directly occasioned by one event. In many cases, the word event is replaced with catastrophe, in others (though less commonly) one originating cause.

The second element is a temporal one, and is present in many (though not all) excess of loss reinsurance contracts. Having defined the ambit of the individual losses eligible for inclusion in the relevant Loss Occurrence, the clause then superimposes a requirement that all such losses must have occurred within a specified period of time. Industry standard periods are 72 hours, 96 hours, 168 hours or up to 504 hours (21 days) in relation to certain perils. Most contracts in fact stipulate a tariff, with different maximum periods for different species of peril. This provision is commonly referred to as the "Hours Clause".

This present case concerned concurrent appeals from two separate arbitration Awards, both concerning excess of loss treaty reinsurance contracts, and each requiring interpretation of both the aggregation language and the Hours Clause. In the first, Covéa Insurance Plc (Covéa) was reinsured by UnipolSai, formerly known as UnipolRe; the second appeal concerned a reinsurance of Markel International (Markel) by General Reinsurance AG. 

The meaning of catastrophe

In both cases the dispute concerned non-damage business interruption losses paid by the reinsureds as a result of closure of their insureds' businesses during the COVID-19 pandemic. Both reinsurance treaties followed the LPO 98 standard wording, under which the reinsureds are required to identify a cohort of individual losses "all…arising out of and directly occasioned by one catastrophe". 

In the case of the Covéa arbitration, the tribunal had concluded that, if and to the extent that the word "catastrophe" called for a "sudden disaster" 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". Consequently, all of the resulting individual losses were eligible for aggregation.

In the Markel arbitration, the tribunal had concluded that the order of the UK Government in March 2020 requiring the relevant premises to close constituted the catastrophe from which all of the business interruption losses arose. In both cases the reinsurers appealed against the respective findings, contending that there was in fact no aggregating catastrophe. 

There is a notable lack of English judicial authority on the meaning of a catastrophe in the context of such contracts, though the matter has been the subject of academic and industry commentary over the years. In their work Reinsurance Law, the authors John Butler and Professor Robert Merkin explain the origin of the LPO 98 wording, noting that the historical aggregating trigger in excess of loss reinsurance contracts had been "any one event". When reinsurers found themselves potentially exposed to aggregated losses suffered over the course of long-term winter freeze conditions they embarked upon a search for an alternative aggregator. Consequently, "the word 'event' was abandoned in favour of the word 'catastrophe' to make it clear that the intention was to cover happenings that were short, sharp and devastating". This was to be contrasted with more protracted events "more accurately described as a state of affairs, such as cold weather".

The subject was also discussed in the work Reinsurance in Practice by the late Robert J, Kiln, himself a co-author of the LPO 98 wording. He observed that, in adopting the wording, the Lloyd's market working party had selected the word "catastrophe":

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

In the present case, reinsurers placed some reliance upon these commentaries. They contended that a progressive spread of a disease lacked the sudden, violent quality necessary to qualify as a catastrophe. So far as concerned the findings of the tribunal in the Markel arbitration, it was also argued that a government order to close premises could not be labelled as a catastrophe, but rather was merely a sensible order in response to the circumstances.

In the Commercial Court, the Judge, Foxton J. accepted that the academic and industry commentaries served as an important reminder of the difference between losses arising from a catastrophe, properly so called, and those which, taken together, might merely be catastrophic in their effect on the reinsured. In the first case, the losses arise from the catastrophe (and so in principle are eligible for aggregation) whereas in the latter case it is the catastrophe which arises from the losses. Clearly, a cohort of disparate losses cannot be aggregated merely because together they might have proved to be catastrophic to a reinsured's balance sheet.

Beyond that, however, the Judge found the materials to be of little assistance. He noted that the commentaries relied upon were historical, far removed from the modern day contracts in the present cases, and authored at a time well before the development of non-damage business interruption insurance. Moreover, whatever may have been the subjective intention of the reinsurers, the contracts did not in fact contain words such as "immediate result" or "sudden, violent physical operation". 

Consequently, it could not be concluded that a catastrophe necessarily called for physical damage, nor was it required to be sudden or violent . All that was required was a "radical discontinuity" with what went before. Put another way, it must be possible to distinguish the period when the catastrophe was in existence from the period when it was not. This necessarily required a sufficient degree of coherence and identifiability to the postulated catastrophe "even if an attempt at a precise temporal delineation would offer scope for legitimate debate and dispute". The Judge also noted that, even if he had found suddenness to be a necessary ingredient of a catastrophe, that would not have assisted in challenging the Award in the Covéa arbitration, since the tribunal in that case had found that, if necessary, the suddenness test was satisfied.

Catastrophe vs. Event

While English case law lacks prior authority on the meaning of a catastrophe, it is replete with cases on the meaning of an event or occurrence ^[2] . Most famously, this includes the "unities" test set out by Lord Mustill in Axa v. Field ^[3] , an "event" being something which, in ordinary speech, "happens at a particular time, at a particular place, in a particular way". In the FCA Test Case ^[4] Lord Hamblen had already concluded that these were qualities absent from the outbreak of COVID-19, albeit in Various Eateries Trading Ltd v Allianz Insurance Plc ^[5] , the Court of Appeal recently accepted that the introduction of COVID-19 into the UK satisfied the criteria of an 'occurrence' (though in that particular case one that was too remote to form a fulcrum of aggregation). The question in the present case was whether the word catastrophe also called for the same unities as those for which the court would customarily search in the case of event/occurrence aggregation.

The Judge concluded that, in so far as the unities test was relevant it should be applied generously in the context of the present treaties. By reference to the Hours Clause (considered further below), he noted that both contracts envisaged that a catastrophe might in some cases persist for a duration of up to 504 hours, and in the case of the Covéa treaty the only geographical limit stipulated in the Hours Clause was "any one country". This necessarily signified that the entire United Kingdom could be the scene of a single catastrophe. The Markel treaty was more restrictive, referring to any one City, but the city in question could still be London, with a metropolitan area of 8,382 km². The Judge also cited with approval the observation in the speech of Lord Briggs, speaking for the minority in the FCA Test Case, that a "hurricane, a storm or flood [all of which the parties agreed could qualify as a catastrophe] …may take place over a substantial period of time, and over an area which changes over time ". 

The Hours Clause

The effect of the Hours Clause is that, even where the reinsured can identify an aggregating catastrophe from which all of the candidate losses arise, only such of those losses as fall within the relevant segment of time will be eligible for aggregation into a single Loss Occurrence for presentation to the treaty. Some treaties will permit only one such Loss Occurrence per event or catastrophe; others will allow the reinsured to present a multiplicity of such Loss Occurrences, each corresponding to a separate segment of time not in excess of the maximum duration specified. Consequently, in the case of a 72 hour limit, where the reinsured has paid claims in respect of physical losses suffered at the hands of a given catastrophe (a hurricane, for example) occurring over a period of (say) 144 hours, it will be required to segregate those losses into two separate Loss Occurrences, each of not more than 72 hours' duration. Each of those Loss Occurrences must, of course, exceed the attachment point in order to give rise to a claim on the reinsurance. The greater the degree of segmentation into separate Loss Occurrences the less likely this threshold will be breached.

Segmentation of physical loss in this way is relatively straightforward, at least in theory. The loss of a given insured's premises suffered in the first 72 hours of a given hurricane belongs in the first Loss Occurrence; a loss paid to an insured whose premises are damaged in the second 72 hour period resides in the second Loss Occurrence, and so on. 

The treatment of business interruption loss is more problematic. Where the reinsured pays for loss of or damage to the insured's property, a significant part of the claim paid to the insured may in fact correspond to the consequential interruption of business suffered in the months following the damage. By definition, this period will extend well outside the specified 72 hour duration. A period of six months' business interruption would equate to no fewer than 60 periods of 72 hours, no one of which is likely to exceed the attachment point. In such cases, the market practice (whether or not correct in principle ^[6] ) is to treat the business interruption as having occurred at the moment of the physical loss, and consequently all subsequent business interruption loss will reside within the same Loss Occurrence as the physical damage from which it resulted, for the purposes of treaty presentation. It is easy to understand why this is so. The damage has happened, and the resulting interruption of business can be said to be a natural and inescapable consequence of it. All that remains is its quantification.

The application to non-damage business interruption is not, however, a matter with which the industry has had to grapple before. If the catastrophe is said to be an outbreak of COVID-19 in March 2020 (or, in the case of the Markel arbitration, the imposition of restrictions at that point in time), this gives rise to the issue whether business interruption suffered several months later is a natural and inescapable consequence of such a catastrophe. Consistent with reinsurers' argument, an interruption of business suffered in, for example, June 2020 was simply the result of something which occurred in June 2020 (namely the prevalence of COVID-19 in June 2020), and did not arise from something which occurred in March 2020. It was impossible to know, in March 2020, whether and for how long the situation would persist, and indeed the legislation by which the restrictions were introduced expressly called for their need to be revisited regularly, in response to developing circumstances.

Unlike in the case of physical damage, reinsurers argued that insureds would only go on suffering interruption of business at the hands of COVID-19 for so long as the disease remained sufficiently prevalent in the community to warrant the continued imposition of the restrictions, a calculation which could only be made on a day-by-day basis.

Faced with these arguments, the tribunal in the Markel arbitration concluded that the individual losses occurred on a daily basis, and so the reinsured could only include payments in respect of closure actually experienced during the stipulated period. The tribunal in the Covéa arbitration reached the opposite conclusion, holding that the losses all occurred when the premises were first closed by order, in March 2020, even though the interruption continued for many months later.

In considering these competing approaches the Judge agreed with the Covéa tribunal, and overturned the decision of the Markel tribunal. He refused to accept that there was a distinction of principle between damage and non-damage business interruption. Such a distinction, he said, was difficult to reconcile with the recent Judgment of Butcher J. in Stonegate Pub Company ltd v. MS Amlin Corporate Member Ltd ^[7] , who held that the "trigger" under the original policies was the forced closure, from which all subsequent business interruption resulted. The Judge in that case held that there would be only one such trigger, which remained operative unless and until the location reopened. 

The Judge in the present case also rejected the idea that, in the conventional physical damage example, all that remained to be determined after the initial "strike" was the pure quantification of the business interruption loss. In fact, contingencies may occur after the strike which could influence the extent of the damage (and hence the extent of the business interruption) but which did not break the chain of causation. In this way, he considered that the distinction between the damage and non-damage scenarios was not as marked as reinsurers suggested.

Finally, the Judge was also troubled by the practical consequences of the "day-by-day" approach. He noted that, under the treaty terms, reinsurers would pay, in addition to the principal of any recoverable losses, the associated expenses and legal costs, and they would derive the benefit of any salvage and recoveries. On reinsurers' case, business interruption and associated expenses would need to be unpicked at the reinsurance level to distinguish between those referrable to interruption experienced during the qualifying hours and that experienced at other times. The problem was all the more acute for the fact that business interruption (and associated costs and expenses) is unlikely to be experienced in a linear fashion; indeed during certain periods the insured may derive a benefit from forced closure.

Summary

This is a significant decision on a number of levels. It is the first English authority on the meaning of the catastrophe aggregator in reinsurance contracts, notwithstanding that the term has been in use in the industry for decades. The court has declined to accept reinsurers' arguments that a catastrophe calls for something short, sharp and devastating and and/or sudden and violent.

Instead, it is enough to establish a radical discontinuity with what went before, distinguishing the period when the postulated catastrophe was in existence from the period when it was not. While the Judge stressed that this is still likely to exclude most if not all attempts to aggregate around a mere state of affairs , it is not easy to see why this is so. If all that is required is a reasonably definable beginning, and an equally identifiable end, and that the intervening period would likely be labelled a catastrophe by the reasonable onlooker, then it is not hard to imagine some very long-term states of affairs that might qualify.

As with many issues arising out of the COVID-19 pandemic, the wording of the reinsurance contract was written at a time when pandemics were not widely contemplated, such that it has not been at all easy to apply the wording to the circumstances of the claims which have emerged out of the pandemic. It remains to be seen whether there is another chapter yet to play out in this interesting case.

[1] Or "Loss" or "Net Loss", depending upon the particular wording used

[2] These two words mean the same thing "unless perchance the contractual context requires some distinction to be made" (Rix J in Kuwait Airways Corporation v. Kuwait Insurance Co [1996] 1 Lloyd's Rep 664)

[3] [1996] 1 WLR 1026

[4] [2021] UKSC1 at 69

[5] [2024] EWCA Civ 10

[6] A matter on which the reinsurers in this case adopted different positions

[7] [2022] EWHC 2548 (Comm)