Covid, Causation, and Coverage: How the Courts are responding to Covid insurance claims

10 May, 2021

An article written by Nigel Eaton QC and Jeremy Brier, Essex Court Chambers

The unprecedented legal restrictions on private and public life over the last year have inflicted untold damage on businesses, particularly in the entertainment, hospitality, and retail sectors. Unsurprisingly, those with business interruption cover have endeavoured to shift the economic pain of repeated lockdowns onto their insurers. The response of the Courts has been a curious combination of the creative and the conservative. On the one hand, the Supreme Court’s approach to causation in Financial Conduct Authority v Arch (UK) Insurance & Others [2021] 2 WLR 123 saw a relaxation of the traditional rules about the nature of the necessary link between the occurrence of an insured peril and the suffering of loss. On the other hand, in both FCA v Arch itself and a number of decisions of the lower Courts, Judges have steadfastly resisted claimants’ attempts to persuade them to adopt a generous approach to construction which would maximise the extent to which Covid-related claims fall within the scope of cover.


FCA v Arch was a test case concerning a sample of 21 insurance policies used in the market, and in relation to which several thousand claims had been made by policyholders for Covid-related business interruption losses. Insurers had refused to pay these claims, on the basis that they were not within the scope of cover. The policies were essentially property damage covers for small and medium sized businesses, which also contained business interruption extensions. They typically required proof that there had been (for example) an outbreak of a notifiable disease either on the insured premises or within a defined proximity, or a prevention or restriction of access to premises following the Government’s order that certain types of business must be closed during the first lockdown.

For example, policies underwritten by Royal & Sun Alliance provided that “We shall indemnify You in respect of interruption or interference with the Business during the Indemnity Period following… any occurrence of a Notifiable Disease within a radius of 25 miles of the Premises.” Other policies followed broadly similar lines, subject to some differences in points of detail, such as the size of the geographical radius. There was no issue that Covid was a Notifiable Disease within the policy. But it was a nationwide – indeed global – pandemic, rather than merely a localised outbreak of the type which the geographical radius wording appeared to contemplate. This raised a question as to how the requirements of causation operated under the policy in circumstances where Covid was rife both within and without the radius, and where the interruption or interference with the business was a consequence of government action which was national in its extent.

It is settled law (eg, Section 55(1) Marine Insurance Act 1906) that, unless a particular policy provides otherwise, an insurer is only liable for losses “proximately caused” by an insured peril. Proximate cause is often equated with “but for” causation, the test being whether or not the loss would still have occurred “but for” the operation of the insured peril: if it would not have done, then the insured peril was a but for/proximate cause; but if the loss would still have happened anyway, then the insured peril was not a but for/proximate cause. Sometimes, “but for” has been referred to as “factual causation”, with proximate cause being termed “legal causation”.

The insurers in FCA v Arch founded their causation arguments on “but for”. Their essential case was that, since the lockdown restrictions were national, the interruption or interference  with any given business would still have occurred even if there had been no cases of Covid at all within the geographical radius. On that basis, the insured peril – ie, occurrences of Covid within the radius – did not qualify as a “but for” cause, and the policies did not respond.

The Supreme Court rejected this argument. It held that the starting point for the causation inquiry in an insurance context is to identify whether an insured peril had any causal involvement in the loss at all. If it did not, it can be disregarded. But if it did, it is then necessary to decide whether the peril made the loss inevitable in the ordinary course of events, in which case, it qualifies as a proximate cause. On this analysis, every single case of Covid in the country qualified as a proximate cause of loss, in the sense that it was an equal cause of the imposition of the national lockdown. Put another way, every case of Covid contributed to the situation in which the business interruption losses were sustained. The upshot was that Covid cases within the geographical radius were every bit as causative as the wider pandemic outside the radius so that, provided there was at least one Covid case within the specified distance of any particular business, the causation requirement was satisfied. The fact that Covid cases outside the radius would have been sufficient in themselves to cause lockdowns was irrelevant.

This marks a real departure from “but for” causation, where the simple (and single)  question is, but for the occurrence of the insured peril (or, in a wider legal context, but for the defendant’s tort or breach of contract), would the loss still have been sustained?

The “but for” test is deeply embedded in the English law of obligations. It has, according to Professor McGregor, “near universal acceptance… the threshold which claimants must cross if their claim for damages is going to get anywhere”. But the Supreme Court considered that it was inadequate to provide a satisfactory answer in the Covid cases. It is, the Court indicated, a blunt instrument. On the one hand, it returns too many false positives in some cases. If, for example, a claimant is run down in a traffic accident, her decision to leave the house that day can be categorised as a “but for” cause: but no-one would normally think that it could be a proximate cause in law. On the other hand, in other cases it excludes situations in which one event would naturally be regarded as a cause of another event. The Supreme Court considered a number of cases where each one of several factors should be considered a proximate cause in law, notwithstanding that they were not “necessary” causes and would not satisfy “but for”. For example, the classic textbook scenarios of two fires converging on a building and burning it to the ground, or of two hunters simultaneously inflicting fatal wounds on a hapless hiker.

Such cases highlight the point, long recognised in the caselaw, that “but for” does not cope well with cases in which there are several causes, each of which is of equal potency. “But for” would suggest that neither fire burns the house, and that neither hunter causes the death of the hiker. But, on any rational analysis, it is obvious that both of them do. However, the textbook examples, and previous decided cases, tend to involve a small number of potential causes: eg, only two in both the fire and the hunter scenarios. At issue in FCA v Arch was whether the same analysis could be extended to scenarios involving hundreds, thousands, or, in the case of the Covid pandemic, even millions of individual causes. The Supreme Court held that it could.

A decision that a single Covid case within a specified radius of a business is a cause in law of losses which flow from a national lockdown triggered by a pandemic involving millions of cases throughout the UK may appear to reduce the requirements of proximate cause to something close to vanishing point. But the Supreme Court considered that the conclusion followed naturally from less extreme scenarios. By way of illustration, the Court considered Professor Stapleton’s examples of 20 individuals who combine to push a bus over a cliff when the efforts of only 14 of them would have sufficed; or of a company’s board voting unanimously to place a dangerous product on the market, when a bare majority vote would have led to exactly the same outcome. Or take another well-worn textbook example, the multiple factories which each pour waste into a river, in circumstances where the effluent from one of them alone would be sufficient to poison the water. Treating “but for” as a minimum threshold which must always be crossed would lead to the absurd conclusion that no individual’s actions caused loss in any of these cases. But the correct legal analysis is surely that the conduct of each individual was causative, and it does not matter that the actions of other individuals would have led to the same loss. The factual situation confronting the Supreme Court in FCA v Arch involved greatly increased numbers of contributing causes, but the same underlying logic applied.

Against this backdrop, the Supreme Court rejected the argument that cases outside the specified radius could be relied upon to negate causation if, taken together, they outweighed the effect of the cases within the radius. The Court concluded that, when giving coverage for notifiable disease, insurers must contemplate that disease may occur both within and outside the radius, and that, in that context, it is unreasonable to attribute to the parties an intention that disease outside the radius should deprive the insured of coverage for an event which results in part from disease within the radius. After all, none of the policies excluded the consequences of events outside the radius (if they had done, the result would presumably have been different: Wayne Tank & Pump Co v Employers Liability Assurance Corp [1974] QB 57). Therefore, the existence of cases outside the radius could not be relied upon to establish that the cases within the radius were not a proximate cause of loss. It was sufficient for a policyholder to show that, at the time of any relevant measure, there was at least one case of Covid within the radius.

This was also the conclusion of the High Court (Commercial) of Ireland in Hyper Trust Ltd v FBD Insurance PLC [2021] IEHC 78. The judgment in that case was delivered shortly after the Supreme Court decision in FCA v Arch, and it refers extensively to the judgments both in the Supreme Court and in the Divisional Court. Interestingly, the Irish Court based its decision that a single case of Covid within the radius was sufficient for causation on the relatively narrow grounds that “following”, the word which was used in both the RSA policy in FCA v Arch and in the Hyper Trust policy to describe the required connection between insured peril and loss, denoted something less than proximate cause or “but for”. This had also been the approach of the Divisional Court in FCA v Arch. The fact that the Supreme Court’s decision was based on a more radical reconsideration of causation raises the question of whether FCA v Arch will herald a wider abandonment of “but for” as the basic litmus test for causation in the English law of obligations. This seems unlikely. The Supreme Court itself certainly did not perceive its decision in revolutionary terms, emphasising that “in the vast majority of insurance cases, indeed in the vast majority of cases in any field of law or ordinary life, if event Y would still have occurred anyway irrespective of the occurrence of a prior event X, then X  cannot be said to have caused Y”. It may therefore be that, in relation to causation, FCA v Arch fittingly represents an extraordinary decision for extraordinary times.


At an earlier stage in the argument in FCA v Arch, the FCA had contended that the conclusion that the policy responded if there was a single Covid case within the geographical radius could be arrived at simply as a matter of construction of the coverage clause, without having to resort to the wider law of causation. The FCA argued that the correct construction was that the policy provided insurance against business interruption resulting from disease wherever the disease occurred, provided that there was at least a single case within the geographical radius. On this analysis, the policy’s reference to disease within the radius was rather unimportant, a purely de minimis requirement which had to be satisfied in order to trigger cover. By contrast, the insurers’ case was that, as a matter of construction, the policy only provided protection against the consequences of instances of disease which occurred within the geographical radius. On this analysis, the radius was the very gist of the coverage, not merely a threshold issue.

At first instance, the Divisional Court accepted the FCA’s argument. The Supreme Court overruled the Divisional Court on this point, although it went on to conclude on a causation analysis (see above) that the occurrence of a single Covid case within the radius was indeed sufficient to engage the policy. Since the Supreme Court ultimately reached exactly the same conclusion as the Divisional Court, albeit by a different route, it may seem academic whether the proper analysis was construction or causation. But the Supreme Court’s decision on construction is significant for the emphasis placed on the “ordinary meaning” of the policy.

The Supreme Court pointed out that the Divisional Court’s approach involved reading any occurrence of a Notifiable Disease within a radius of 25 miles of the Premises” as if it said, “a Notifiable Disease of which there is any occurrence within a radius of 25 miles of the Premises.” It chastised the lower Court for neglecting to explain “how, as a matter of the English language”, this transposition was possible. The Divisional Court might, perhaps, have cited the scepticism of Lord Hoffmann, who, in Investors Compensation Scheme v West Bromwich Building Society [1998] 1 WLR 896, famously queried whether words have an “ordinary meaning” in day-to-day discourse (or in contracts). He argued that any human utterance necessarily takes meaning from its concrete context, not from abstract definitions. But the Supreme Court relegated Lord Hoffmann’s approach, which permits “verbal rearrangement or correction of the words used”, to “those rare situations in which the court can be satisfied that… something must have gone wrong with the language”. It said that, if contractual words have an “obvious meaning” as a matter of the English language, that meaning must be applied.

Moreover, the Supreme Court emphasised the contribution to contractual construction not of Lord Hoffmann, but of his equally eminent contemporary, Lord Mustill. In particular, the Court highlighted Lord Mustill’s “cautionary words” that, to force upon a clause a meaning which it cannot sensibly bear is to engage in the illegitimate exercise of substituting for the contract which the parties actually made a contract which the Judge thinks they could better have made. This aspect of the judgment appears to re-affirm (following on from Arnold v Britton [2015]  AC 1619 and Wood v Capita [2017] AC 1173) that the high-water mark of ordinary language scepticism among the judiciary in England & Wales has passed, at least for now

Construction arguments appealing to “commercial common sense” (or, perhaps more accurately, to a generalised sense of “fairness”) have met with a similar lack of success in several first-instance cases concerning Covid and insurance.

The policy in Rockliffe Hall v Travelers Insurance Co Ltd [2021] EWHC 412 (Comm) was, like those in FCA v Arch, in a traditional form, with business interruption cover provided as an add-on to what was essentially insurance against property damage. The insured property in question was a five-star luxury hotel and resort in County Durham, which doubled as a lucrative conference and wedding venue. The business interruption extension applied to “loss directly resulting from interruption to or interference with the Business in consequence of an outbreak of Infectious Disease within 10 miles of the Business Premises”. Structurally, the cover was very similar to the RSA policy analysed in FCA v Arch. But there was a crucial distinction. The RSA policy responded to “Notifiable Disease”, which was defined as any infectious disease declared notifiable by the authorities. The scope of cover therefore had a dynamic aspect, which was capable of responding to newly emergent infections (including Covid) which only became notifiable after inception of the policy. By contrast “Infectious Disease” in the Rockliffe Hall policy was defined statically, by reference to a closed list of 34 specified illnesses and conditions. As it happened, that list represented the range of conditions which had been declared notifiable down to 2010. But, because the list was closed, it was incapable of extending to conditions, such as Covid, which had only emerged since then.

In these circumstances, the claimant was compelled to contend that, although Covid was a new disease, it could nevertheless be brought within one or more of the 34 pre-existing notifiable conditions listed in the definition. The Claimant based its case principally on “plague”, arguing that this was a generic reference to any infectious disease with high mortality. The Court concluded that the fatal difficulty with this argument was that, if “plague” bore a generic meaning, then it would embrace virtually all of the other listed illnesses and conditions, which would render the greater part of the (carefully-structured) definition of “Infectious Disease” redundant. The Judge held that the obvious and common-sense construction was that “plague”, just like “cholera”, “smallpox”, and “typhus”, which appeared next to it in the definition, referred to a specific condition (viz, the y pestis pathogen and its consequences).

The claimant’s backstop argument that Covid could be equated with “encephalitis”, or “meningitis”, or both, because there was some evidence that it could cause these diseases, received similarly short shrift. As the Judge pithily put it, the fact that Covid might cause encephalitis or meningitis did not mean that it was the same thing as encephalitis or meningitis, and so could not bring it within the scope of cover which insured against those diseases.

The High Court (Commercial) of Ireland was equally unpersuaded that Covid was really something else in Brushfield v Arachas [2021] IEHC 263. This was a claim by a Dublin hotel and bar under a policy which was issued by AXA, and so presumably likely to have been in a relatively common form. As in Rockliffe Hall, the disease aspect of the business interruption cover was limited by reference to a closed list which did not include Covid. The Court rejected the claimant’s argument that a claim for Covid-related losses could be hung on “encephalitis”, one of the named diseases. In an alternative argument which has not yet been tested in the Courts in England & Wales, the insured suggested that the enforced closure of the business was the consequence of an order of a public authority resulting from “a defect in the drains or other sanitary arrangements at the premises”, within the policy’s “Murder, Suicide or Disease” clause. The basis for this imaginative argument was that it was impossible, within the physical constraints of the hotel and bar environment, to maintain sufficient levels of social distancing for safe operation. This suggestion that practical constraints on social distancing constituted a “defect in sanitary arrangements” was clearly something of a stretch, and the Court was not convinced. The fact that the claimant was constrained to fall back on this point illustrates the difficulties which insureds may face in bringing Covid claims within the scope of cover under policies which were drafted before “lockdown” entered everyday language.

Similarly, in TKC London Ltd v Allianz Insurance PLC [2020] EWHC 2710 (Comm), the Commercial Court concluded that the policy wording was simply not appropriate to extend to Covid losses sustained by a restaurant. In that case, the policy did not refer to illness or disease at all, and, on an ordinary language construction, it insured only against business interruption which was consequential upon physical destruction of or damage to the insured business premises. The claimant put its case on the basis that “accidental loss of or destruction of or damage to property used by the Insured at the Premises for the purpose of the Business” was capable of extending to loss of the ability to use the premises as a result of lockdown, notwithstanding that the premises remained physically intact. But the Court had little difficulty in concluding that the insuring clause did not extend to transient deprivation of use of the premises. The claimant’s alternative argument that the policy was engaged because foodstocks which had been stored on the premises had deteriorated during lockdown also failed. Even on the hypothesis (which the Court thought was wrong in any event) that this natural deterioration constituted “accidental… damage to property used by the Insured… for the purposes of the Business”, it was a consequence of the business interruption, not the cause of it.    

In a rather different context, two recent commercial landlord and tenant cases also show the Courts resisting expansive arguments as to the scope and significance of insurance cover in relation to Covid-related losses, albeit that business interruption insurance was only of indirect significance.

In Bank of New York Mellon v Cine-UK Limited [2021] EWHC 1013 (QB), the issue was whether tenants of commercial premises remained liable under their leases to pay their rents in full notwithstanding that they had been compelled to close their businesses during lockdown. The tenants in question were Cine-UK, Mecca Bingo and SportsDirect, three well-known and substantial commercial entities heavily affected by the Covid regulations. The issues in the case centred on Rent Cesser Clauses which provided, in essence, that if the tenants’ use of or access to the premises was affected by specified risks, payment of rent would be suspended. The leases provided that the landlords would insure against the specified risks, with the tenants paying the premium. The broad commercial intention was that the landlords would look to the insurers, not to the tenants, if any of the specified risks occurred and caused a loss of rent. The tenants argued that the Covid pandemic and lockdown were within the scope of the specified risks and of the landlords’ insurance, and that the Rent Cesser Clauses were therefore engaged. If this was right, then the risk of Covid-related losses fell on the insurers.

The argument failed. The Court held that, rather as in TKC London Ltd v Allianz Insurance PLC (see above), the focus of the contractual provisions was on physical loss of or damage to the premises. Since there had been no such loss or damage, the Rent Cesser Clauses were not engaged, and the tenants remained liable to pay. It followed that the landlords had no claim under the insurance policies. Although the policies included disease as an insured peril, they provided insurance against loss of rent, and the landlords had suffered no loss of rent in circumstances where the Rent Cesser Clauses did not apply.

Commerz Real v TFS Stores [2021] EWHC 863 raised similar issues. The claimant, the leasehold owner of the Westfield retail complex, sub-let premises to the defendant, which traded as “The Fragrance Shop”. Again, the issue was whether the tenant remained liable for rent during lockdown, or whether the landlord should look to its insurance. In this case, the tenant argued that the landlord was in breach of its insuring obligations. But the Court accepted the landlord’s case that this was irrelevant because, absent physical loss or damage, the Rent Cesser Clause was not engaged, and so the landlord had not suffered any loss which would have been covered by insurance in any event. As in the Bank of New York Mellon case, the Court effectively concluded that the tenant’s argument that the Rent Cesser Clause was engaged because disease was (or should have been) covered by the landlord’s insurance put the case the wrong way around. The correct starting question was whether the Rent Cesser Clause applied. If it did not, it followed that there was no insured loss to which the insurance responded.

In neither of these landlord and tenant cases, therefore, did the landlords’ business interruption insurance policies provide any financial assistance to the tenants in relation to their lockdown losses. The moral, according to the Court in Bank of New York Mellon was that, if the tenants had wanted the security of protection against the purely economic consequences of a pandemic, they could and should have taken out business interruption insurance for their own benefit. However, as Rockliffe Hall v Travelers Insurance Co Ltd  and TKC London Ltd v Allianz Insurance PLC demonstrate, whether that would have put them in any better position would have very much depended on the particular policy wording, and by no means all business interruption policies are sure to respond to the unprecedented circumstances of the pandemic.

Procedural Aspects

Finally, the commendable speed and efficiency with which the Courts have managed to dispose of Covid-related insurance cases is worthy of note.

FCA v Arch was brought as a test case under the Financial Markets Test Case Scheme (now embodied in CPR Practice Direction 63AA.6). The Scheme enables the Court to rule on “issues of general public importance on which immediately relevant authoritative English law guidance is needed”, without the need for any present cause of action between the parties. As it happened, the FCA and the insurers signed a Framework Agreement jointly agreeing to the litigation and making proposals (considered by the Divisional Court at the first CMC) for how the hearing should be conducted, eg, by a two Judge Divisional Court, without costs risks etc. The case was argued before the Divisional Court for 8 days in late July 2020, and the (165 page) Judgment was delivered in September. Permission to appeal direct to the Supreme Court was given in November 2020. The appeal was argued for 4 days later in the same month, and Judgment was delivered in January 2021.

The four other England & Wales cases mentioned above were each decided on applications for summary judgment/reverse summary judgment and/or to strike out the claims. In each of them, the insured or tenants argued that there were questions of complexity and importance which were not suitable for summary determination, but the Court took the robust view that it had before it such (limited) evidence as was likely to be relevant, and that the key points were essentially matters of law or construction which could safely be determined on the basis of submissions, without a full trial. The two Irish cases were not disposed of summarily, but the very full judgment in each was handed down within an impressively short period after trial.

It is fitting that these cases were resolved in this business-like fashion during the legal year which marked the 125th anniversary of the creation of the Commercial Court of England & Wales. When Sir James Charles Mathew, the first Commercial Judge, began to exert a grip on the conduct of High Court commercial litigation after nearly a generation of aimless drift, he was determined to satisfy demands from the commercial community that their disputes should be resolved promptly. One of his favoured techniques was to identify at an early stage any points of law or construction on which a case hinged, and decide them summarily. Mathew, who was a proud Irishman, would have been satisfied to see his methods and his judicial philosophy alive and well, on both sides of the Irish Sea, a century and a quarter after he first demonstrated their value.

Written by Nigel Eaton and Jeremy Brier, Essex Court Chambers

This note is provided free of charge as a matter of information only. It is not intended to constitute, nor should it be relied upon as constituting, legal advice, and no responsibility is assumed in relation to the accuracy of the contents of the same as regards anyone choosing to rely upon it.