Business interruption insurance and COVID-19

Thursday, September 17, 2020

In the United Kingdom, the Financial Conduct Authority (FCA) brought a test case on behalf of insurance policy holders against a group of six insurers to obtain clarity as to the operation of business interruption insurance for claims arising out of the COVID-19 lockdown.

The circumstances of the pandemic and the policy wordings are common to New Zealand. 

Background to the FCA judgment

Insurance cover for business interruption losses is typically triggered by material damage giving rise to such losses.

Although it has been suggested in some US courts that the COVID-19 virus might cause material damage to property, the accepted definition of damage – physical alteration to the integrity of property – removes any serious possibility of that approach succeeding elsewhere in the common law world. Accordingly, if there is to be cover for BI losses in a pandemic, there has to be some trigger of cover independent of physical damage. Such triggers do exist in policy extensions for disease and loss of access. Many COVID-19 claims under extensions were rejected by UK insurers, leading to an arrangement between the market and the insurance regulator whereby the most important wordings would be construed by way of guidance to the market for future settlements.

The decision was handed down on Tuesday 15 September Financial Conduct Authority v Arch Insurance and Others [2020] EWHC 2448 (Comm). The English High Court, consisting of Lord Justice Flaux and Mr Justice Butcher reviewed 21 policies, as test cases for 700 policies issued by over 60 different insurers and affecting up to 370,000 policyholders. The judgment is close to 600 paragraphs in length. The guidance given by the High Court will be of direct relevance to insurers and brokers in New Zealand, as the wordings under consideration are identical or at least very similar to those found in the NZ market. It should be stressed that the judgment does not deal with individual claims.

By way of factual background, it suffices to say that COVID-19 became notifiable in the United Kingdom on 5 March 2020 and, following the publication of a series of voluntary guidelines legal steps were taken to impose a lockdown in measures beginning on 21 March 2020.

It is not possible here to examine each of the 21 wordings under review, and this note sets out the basic principles laid down by the Court.

The “counterfactual”

Key to the judgment is the Court’s approach to the “counterfactual”.

BI policies by their insuring terms and by the “trends” clause operate on the basis that the assured can recover financial losses caused by the relevant insured peril for an indemnity period as laid down by the policy. To ascertain the amount of that loss it is necessary to work out what the assured’s revenue would have been had the insured event not occurred. That is typically done by looking backwards to see how the business had performed before the occurrence of the insured peril. Sometimes a fixed period – eg, a year – is laid down to provide a comparator, in other cases there is no period. But in most if not all cases there is a “trends” clause or its equivalent. Its purpose is to allow the Court to take account of exceptional events that may have depressed or increased revenue in the earlier comparator, and also to take account of anticipated exceptional events in the indemnity period.

To determine what would have happened but for the insured peril, it is necessary to assume that the insured peril did not occur: that is the counterfactual. Based on the actual situation in which there was an outbreak of disease followed by Government lockdown measures, and policies responding to Government measures, the Court was faced with two competing counterfactuals. The first was to assume the existence of the COVID-19 pandemic but in a scenario where no Government measures had been taken in respect of insured premises. Inevitably, with people voluntarily choosing to adopt a policy of social distancing or being forced to through illness, there would have been a serious impact on retailing, manufacturing and the provision of services. Using this counterfactual, the measure of lost revenue resulting from the lockdown would be the difference between an economy hit by COVID-19 and an economy subject to Government lockdown. That approach was said to measure the actual effect of the lockdown measures.

The alternative counterfactual was to assume that there had been no pandemic at all. The measure of lost revenue would then be the difference between what would have been earned a fully-functioning economy free of disease measured against what was actually earned following the lockdown. Plainly the latter gives a far greater recovery.

The Court adopted the latter. It did so by holding that the starting point was to indemnify the trigger for cover, ie, the insured peril. The insured peril was not Government lockdown leading to restricted access to or closure of premises, but rather COVID-19 resulting in Government lockdown. In other words, by treating COVID-19 as part of the insured peril, the counterfactual had to disregard COVID-19 itself. That reasoning removed the need to consider causation issues. The approach is to define all of the elements that constitute the insured peril (pandemic, lockdown, restricted or excluded access) and to “strip out” those elements from the counterfactual.

One important consequence of this reasoning was the rejection by the Court of the reasoning in the much-discussed case Orient-Express Hotels Ltd v Assicurazioni Generali SpA [2011] Lloyd’s Rep IR 531. There, the assured’s hotel in New Orleans was damaged by Hurricanes Katrina and Rita, and was forced to close for repairs. At the same time the City was put under lockdown and visitors were unable to come in. The owner sought a BI indemnity, but it was effectively denied by the Court holding that the correct counterfactual was an undamaged hotel in a damaged city, so that it could not be said that the BI loss flowed from the hurricanes.

It is clear after the FCA case that in ascertaining the counterfactual it is necessary to disregard the hurricanes as they are part of the insured peril, so that the correct counterfactual is a city unaffected by hurricanes, allowing the assured to recover BI losses for the indemnity period after the hurricanes had inflicted material damage.

Local and national issues

All of the relevant BI wordings required interruption to business at the insured premises. In addition, they all tied cover to something happening either within a specified radius of the premises or in the vicinity of the premises. For example, the wording might cover “interruption of 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.” Similar wording might refer to “vicinity” rather than a specific radius and also talk of the threat of infection in the relevant locality. The difficulty is that a pandemic does not adhere to geographical rules.

The Court addressed this problem by holding that if there is a generalised outbreak of the disease, as long as there is one instance of it in the locality then there is cover, and it does not matter that the interruption is the result of the general incidence of the disease. In other words, there is no need for a causal connection between the local occurrence and the interruption of business. Equally, if there is cover for government action in the locality, national government action necessarily implies action in the locality and so there is coverage.

However, care must here be taken. Compare the above wording with this formulation: “Loss resulting from interruption of or interference with the business in consequence of the following events” namely “any occurrence of a notifiable disease within a radius of 25 miles of the premises.”  This wording is quite different. It uses the language of “event” (or, it could be “incident”) which points to something happening at a particular time, in a particular place and in a particular way. Similarly, the requirement that interruption is the result of something happening in the locality of the premises – whether an occurrence, manifestation or threat of disease – means that the policy responds only to a local event and not to a national one even though there are national implications.

So, the distinction is clear in principle. Policies expressed in general terms and referring to the locality will provide cover even though there are national implications or measures, as long as a single case can be proven. By contrast, if the cover is worded in specific terms requiring a local event, the closure or loss of business must be by reason of that event alone.

Meaning of specific words

A number of words commonly used in business interruption policies fell to be considered. These were as follows:

  • First, there is a distinction between the “occurrence” of a disease and the “manifestation” of a disease. An “occurrence” requires actual infection, whether or not manifest or diagnosed, whereas a “manifestation” requires either symptoms or diagnosis.
  • Secondly, in the context of access, a distinction is to be drawn between “prevention” of access or use and “restriction” of access or use. The word “prevention” means a complete ban on access, whereas a “restriction” is a hindrance to access. However, the prevention must relate to the insured business, and so it is necessary to examine that business to see whether it is prevented or restricted. By way of example, if the assured is a restaurant, the closure of the restaurant amounts to prevention of the insured business, and the fact that the restaurant is then able to switch its activities into takeaway only does not preclude a finding of “prevention” because the previous business cannot continue.
  • Thirdly, “interruption” does not require a complete closure of the business.
  • Fourthly, as regards intervention by public authorities, the Court ruled that a “restriction imposed” refers to something enforceable by law and not merely advisory.

Finally, the Court was in some cases faced with policy coverage under an extension that was all but eliminated by exclusions. The Court regarded such an interpretation of the exclusions as absurd and refused to permit them to nullify specific extensions.

Proof of COVID-19

Given that the Court’s interpretation of the policies requires at least one case of COVID-19 in the locality of the insured premises, the question then arises as to how such proof is possible. There are of course statistics governed by the authorities, but they may not be fully reliable. The Court instead relied upon the decision in Equitas Ltd v R&Q Reinsurance Company (UK) Ltd [2009] EWHC 2787 (Comm) for the proposition that modelling could be used. In Equitas the impossibility of establishing losses within the LMX Spiral led Gross J to accept proof by actuarial modelling, accepting the least favourable model to the claimant. The Court in FCA was not required to adopt any specific modelling mechanism, as insurers had conceded that distribution based on an “undercounting” analysis could be used to discharge the assured’s burden of proof.

For further information please contact Rob Merkin or a member of our insurance or litigation and dispute resolution team.


Disclaimer: the content of this article is general in nature and not intended as a substitute for specific professional advice on any matter and should not be relied upon for that purpose.


Share this publication