March 3, 2022

In Corbin & King v AXA, Mrs Justice Cockerill found policyholders were entitled to an indemnity in respect of business interruption losses sustained during the Covid-19 pandemic. The case was heard on an expedited basis, with the judgement providing much-anticipated guidance in respect of Prevention of Access clauses in business insurance policies in light of last year’s business interruption insurance test case, FCA v Arch [2020] EWHC 2448 (Comm), [2021] UKSC 1.

The decision is considered by Carola Binney of 4 New Square.

Miles Harris of 4 New Square was instructed as junior counsel for Axa.


The Claimants (“the Policyholders”) were the owners of a number of well-known London restaurants, including The Delaunay and The Wolseley. The Policyholders took out business insurance underwritten by the Defendant (“Axa”), pursuant to a policy covering the period from November 2019 to November 2020 (“the Policy”).

The business interpretation cover insured gross profit, additional working costs and tips and service charges across the Policyholders’ premises. It included a Prevention of Access clause, which read:

“Denial of access (non-damage) cover

We will cover you for any loss insured by this section resulting from interruption or interference with the business where access to your premises is restricted or hindered for more than the franchise period shown in your schedule arising directly from:

1 the actions taken by the police or any other statutory body in response to a danger or disturbance at your premises or within a 1 mile radius of your premises.


We will not cover you where access to your premises is restricted or hindered as a result of


  1. notifiable diseases as detailed in the Murder, suicide or disease cover…”

Covid-19 was not among the notifiable diseases listed in the “Murder, suicide or disease” clause. The “franchise period” after which cover could kick-in was two hours, and the maximum indemnity period was 12 weeks.

The Schedule accompanying the full Policy wording included a financial limit in respect of Prevention of Access cover, stated to be “100% of the sum insured or £250,000 whichever is less”.

The Policy was issued on 7 February 2020. The impact of subsequent events on our ability to dine at the Policyholders’ establishments is a topic of which readers will need little reminding. A fortnight after the Policy was issued, the first set of lockdown restrictions required restaurants to close or move to a takeaway-only service. Having reopened on 4 July 2020, business was again limited by an enforced 10 pm closing time introduced on 24 September 2020 – before ceasing entirely with the imposition of new lockdown measures on 5 November 2020.

The Policyholders sought to claim in respect of their ensuing losses; in late November 2020, Axa declined cover.


The case is the first time the Commercial Court has considered the implications of the Supreme Court’s decision in FCA v Arch for unappealed aspects of the Divisional Court’s decision in that case. The facts of the test case litigation will be as familiar to many readers as the restrictions on dining at The Wolseley: 21 “lead” business interruption insurance policy wordings were considered first by a two-judge Divisional Court and then by the Supreme Court.

The FCA v Arch litigation was very largely concerned with “disease clauses”, being those triggered by the occurrence of a “notifiable disease” within a certain radius of the insured premises. The Supreme Court upheld the Divisional Court’s finding that the disease clauses considered generally did provide cover in respect of the Covid-19 pandemic.

There was no appeal, however, in respect of the first instance ruling that the various prevention of access clauses (including Non-Damage Denial of Access (“NDDA”) and “action of competent authority” (“AOCA”) clauses) considered alongside the disease wordings generally did not provide effective cover in respect of the pandemic.

The policy wordings considered by the Divisional Court that most closely resembled Axa’s were AOCA clauses in polices written by MS Amlin and Zurich. The relevant MS Amlin clause (labelled MSA 1 in the FCA v Arch litigation) provided:

“We will pay you for:

  1. Action of competent authorities

loss resulting from interruption or interference with the business following action by the police or other competent local, civil, or military authority following a danger or disturbance in the vicinity of the premises where access will be prevented….”

The Zurich clause provided:

“Action of competent authorities

Action by the police or other competent local, civil or military authority following a danger or disturbance in the vicinity of the premises whereby access thereto will be prevented provided there will be no liability under this section of this extension for loss resulting from interruption of the business during the first 3 hours of the indemnity period.

The maximum indemnity period is 12 months.”

Like the Axa wording considered by Cockerill J, these clauses both respond to a “danger or disturbance”. Other clauses considered by the Divisional Court in FCA v Arch referred to an “emergency” or an “incident”, which Cockerill J held (at [156]) gave rise to a different construction: she made reference to the Divisional Court’s conclusion at [404] of its judgement that the word “incident” “should be given the same essential meaning as “an event”: something which happens at a particular time, at a particular place, in a particular way.”

Neither the MS Amlin clause nor its Zurich counterpart was identical to the Axa clause. Neither contained a disease exclusion, and both referred to dangers or disturbances in the “vicinity” of the insured premises rather than within a specified radius (although Cockerill J considered that the latter point made no material difference to the correct construction). The list of competent authorities was also differently expressed in the MS Amlin and Zurich clauses, which made specific reference to a “local” authority (as opposed to Axa’s reference to “the police or any other statutory body”).

Nonetheless, the Divisional Court made findings about the scope of the “danger or disturbance” wording on which Axa relied before Cockerill J. In particular, the Divisional Court held at [436], with respect to the MS Amlin clause:

“the cover under this clause is a narrow, localised form of cover. The paradigm example of what it covers, as [Mr Kealey QC] submitted, is the bomb scare or gas leak in the vicinity or neighbourhood of the premises which causes the authorities (whether the police or the army in those examples), exercising statutory powers, to evacuate insured premises and require policyholders and their employees and customers not to access the premises.…it follows that, on the true construction of the AOCA clause, the government action in imposing the Regulations in response to the national pandemic cannot be said to be following a danger in the vicinity, in the sense of in the neighbourhood, of the insured premises.”

The Divisional Court therefore concluded that cover would only be available in respect of losses caused by the pandemic if the policyholder was able to show that “it was the risk of COVID-19 in the vicinity, in that sense of the neighbourhood, of the insured premises, as opposed to in the country as a whole, which led to the action of the government in imposing the Regulations” (at [437]). Similar remarks were made with respect to the Zurich wording at [499]-[501].

That analysis was not challenged on appeal. The Supreme Court did, however, provide extensive analysis of the causation requirements in respect of the disease clauses. The majority considered, at [176], that each case of Covid-19 was a concurrent cause of the restrictions imposed by the government. While the disease clauses only responded to the effects of infections occurring within the specified radius of the insured premises, each local case could therefore be properly considered to have caused the national restrictions (and therefore the relevant business interruption). At [191] of the Supreme Court’s judgement:

“…there is nothing in principle or in the concept of causation which precludes an insured peril that in combination with many other similar uninsured events brings about a loss with a sufficient degree of inevitability from being regarded as a cause – indeed as a proximate cause – of the loss, even if the occurrence of the insured peril is neither necessary nor sufficient to bring about the loss by itself.”

In Corbin & King v AXA, the key issue between the parties with respect to coverage was whether the Divisional Court’s approach to similarly worded Prevention of Access clauses remained correct in light of the Supreme Court’s analysis of causation.

The Policyholders argued that they were entitled to an indemnity provided they could establish that (a) there were cases of Covid-19, or a threat of such cases, within a one-mile radius of each insured restaurant (said to be a “danger” for the purposes of the Prevention of Access clause) and (b) each such local case or risk was, combined with actual and threatened cases elsewhere in the UK, an effective cause of the restrictions that led to the prevention of access to the premises.

Axa argued that the Divisional Court in FCA v Arch had been correct to find that the similarly worded MS Amlin and Zurich clauses provided only narrow and localised cover in respect of a “danger or disturbance” specific to the locality of the insured premises. On that basis, Axa argued that the Policyholders were not entitled to an indemnity, because they could not (as they accepted) demonstrate that the denial of access was specifically caused by the risk of Covid-19 within a one-mile radius of each insured restaurant.


Cockerill J found for the Policyholders. She considered that although a decision of the Divisional Court took precedence over that of the High Court, she was not bound for two reasons. First, the wording before her was materially different to that considered by the Divisional Court. Second, she took the view that the argument had been framed before her in a sufficiently different way that she was not bound to reach the same outcome.

In FCA v Arch, the policyholders argued that the danger was the pandemic itself: Cockerill J summarised the argument as having been that “although COVID was everywhere, because it was everywhere it was also local, hence it was in the vicinity or radius” (at [163]).

By contrast, in Corbin & King, the nub of the Policyholders’ argument on coverage was that a single case of Covid-19 within a one-mile radius was “a danger” – and moreover sufficient (on the Supreme Court’s analysis) to have caused the national restrictions. While the cover might be narrow and localised, it could nonetheless respond to the consequences of the national lockdown (as “caused” by localised cases). This line of analysis had not been explored before the Divisional Court, such that the Supreme Court’s decision had “moved the goalposts”.

Cockerill J therefore proceeded to construe the Prevention of Access clause in question, without considering herself bound by FCA v Arch to find that no cover existed. She emphasised the importance of approaching the clause with reference to how the words used would be understood by a reasonable policyholder. Such a policyholder can be expected to have read the policy conscientiously, but cannot be expected to subject its terms to minute textual analysis with reference to, for example, evidence of the historical derivation of particular wordings: “we approach the Policy on its words, as if we are a small businessman, albeit with a broker to assist us” (at [180]).

Applying that approach, “Danger” was held to be a generic word, not constrained to transient incidents posing a health and safety risk. The clause did not bite until the danger had existed for at least two hours and the indemnity period was up to 12 weeks; while the indemnity period could not be conflated with the currency of the danger, it was nonetheless relevant in indicating that it was envisaged that the consequences of the danger might last for more than 12 weeks (hence the need for a temporal limitation on cover). This was held to undermine Axa’s submission that the danger needed to be transient.

In terms of the clause’s geographic scope, the reference to “at the premises” and the one-mile radius required a “local link”, but did not demand that the danger must be something of very local significance (of which a bomb scare or a gas leak had been proffered as classic examples).

On this basis, Cockerill J concluded that, “although not the most obvious fit”, “danger” was capable of extending to disease (at [181]). In respect of the radius requirement, she concluded that while it was necessary for the danger to have had a local manifestation, it did not need to be exclusively local. This conclusion was strengthened by the description of the competent authority: “any other statutory body” was wide enough to encompass the central government.

As to the express exclusion of certain notifiable diseases from cover, this was ultimately held to assist rather than hinder the Policyholders’ case. Axa argued that a reasonable person reading the “Murder, suicide or disease” (“MSD”) clause and the Prevention of Acccess clause together would conclude that the Policy was intended to provide cover only in respect of the closed list of diseases included in the MSD clause. Cockerill J disagreed, taking the view that the “logical correlate” of the exclusion of some diseases was that other diseases could in theory be covered (at [187]).

Having reached that conclusion on construction, Cockerill J then applied the Supreme Court’s approach to causation to find that the Policy responded to the losses suffered as a result of the national lockdown measures.


Cockerill J then went on to consider a further issue in relation to quantum, regarding the correct approach to the Prevention of Access clause’s cover limit of “100% of the sum insured or £250,000 whichever is less”.

It was agreed that the limit could be applied afresh in respect of each of the three sets of restrictions imposed during the policy period (being the two lockdowns and the enforced 10 pm closing time). The issue between the parties was whether the limit was £250,000 per restriction (i.e. £750,000 in total) or £250,000 per restriction per premises.

Axa argued that the business interpretation provisions in the Policy did not, unlike those in respect of property damage, stipulate separate sums insured specific to particular restaurants; it was said that the use of an aggregate limit suggested that it was intended to apply to all the premises insured.

The Policyholders stressed that each restaurant was a separate business, and as such the interruption to the business of each one gave rise to a separate claim under the Policy, which provided cover in relation to “interruption and interference with the business where access to your premises is restricted or hindered.”

Cockerill J again found for the Claimants. Referring to the above-quoted policy wording, she held at [239]:

The premises were in different locations and could well be differently affected by a danger triggering cover…The word “premises” points to each restaurant/café and that distinction illuminates how a separation of interests may well operate – and that in turn points to separate limits.


2021 was not a good year for insurers, and the decision in Corbin & King v Axa suggests that the tide has not yet turned.

Cockerill J’s judgement was ultimately grounded in issues of construction, and as such there remains scope for insurers with differently worded Prevention of Access clauses to argue that an approach more akin to that taken by the Divisional Court in FCA v Arch ought to apply. Cockerill J’s judgement indicates, for example, that clauses in which the competent authority is described in more expressly local terms may well not extend to the effects of the national lockdown. Insurers with clauses referring to an “incident”, rather than a “danger or disturbance”, also appear to stand a better chance of successfully declining cover.

Nonetheless, the decision is a further favourable development for policyholders and will do little to soothe insurers’ nerves ahead of the further linked business interruption insurance trials scheduled to be heard later this year – Stonegate Pub Company v MS Amlin is next on the list, with an expedited trial of certain issues set to commence in June 2022.

We are all tired of hearing that the Covid-19 pandemic was “unprecedented”, but Cockerill J’s judgement is, more generally, the latest illustration of the difficulties in seeking to apply contractual terms to events that were likely completely unimaginable to either party at the time those terms were drafted.

While characteristically rigorous and considered, the judgement is not light reading. The insurance contracts of the future will doubtless include provisions tailored to the risk of pandemics, but those of the pre-Covid world cannot be made to translate to the realities of the last two years without some particularly headache-inducing legal analysis.

As Cockerill J herself acknowledged in respect of her conclusions on the construction issue, at [204]:

“it might be said to be strange that if this is the correct construction, that it has taken at least three iterations of the argument to reach this point. However, it is not unknown for construction arguments to have to be iterated before a satisfactory conclusion is reached.”

With several more linked cases (and possible appeals) in the Commercial Court pipeline, it remains to be seen whether the iteration will stop there.

Disclaimer: This article is not to be relied upon as legal advice. The circumstances of each case differ and legal advice specific to the individual case should always be sought.

© Carola Binney 4 New Square, 3 March 2022