CORBIN & KING LIMITED v AXA INSURANCE UK PLC

This case concerns the scope of cover provided by a Denial of Access (Non Damage) (“NDDA”) clause in a combined business insurance policy issued by the Defendant insurer pursuant to a “Business Combined Insurance policy” which commenced on 12 November 2019 and expired on 11 November 2020 (“the Policy”).

The Policy was issued to the Claimants, the owners and operators of a number of well-known restaurants, cafes and other establishments in and around London.

The two key issues in the case are:

Whether the NDDA clause provided effective cover for loss resulting from restrictions on access to the Claimants’ premises under government regulations passed in response to the COVID-19 pandemic in the course of 2020 (“the Coverage Issue”).

Whether, if the NDDA clause did provide cover, there was a single limit of £250,000 in respect of all premises for any one claim, or whether there was a limit of £250,000 for each set of premises (“the Quantum Issue”).

The claim thus concerns COVID business interruption insurance, and was entered on this Court's COVID BI list. Determination of these issues has been expedited in the light of the fact that it appears that the issues in this case affect not just these litigants but a considerable range of other businesses, and that a number of disputes may well be short circuited by an early determination of the issues.

In particular the solicitors instructed for the Claimants represent a number of other parties with materially identical NDDA wordings, and produced evidence of a significant number of other parties, both those who had already obtained legal representation and those who had not, who were in a similar position. All of those parties indicated that they were prepared to await the outcome of an expedited hearing and would not wish to make separate representations on the issues.

As will be set out below, the parties’ submissions on the Coverage Issue turn in significant part on analysis of the reasoning of the Divisional Court and the Supreme Court in The Financial Conduct Authority v Arch Insurance (UK) Limited and others [2020] EWHC 2448 and [2021] UKSC 1 (“Arch”).

The court is not being asked to determine loss adjustment issues, which will be addressed by the parties if the Claimants succeed in this action.

The Policy was issued on 7 February 2020, with the period of insurance stated to be from 12 November 2019 to 11 November 2020. The Policy is made up of two documents: the Schedule and the Full Policy Wording.

The Schedule is subtitled “Your Business Combined Insurance Policy”. Under “Your details” on the first page of the Schedule, “The insured” is stated to be “Corbin & King Limited & Subsidiaries”. The Claimants’ names are found on the second page of the Schedule under the heading “Additional policy information” and the sub-heading “Full Client Name”. They are set out as follows:

In the “Property damage section” which follows, twelve different restaurants and/or addresses are then listed separately as numbered “premises”. With two exceptions the premises are all in London.

The “Business interruption section” then begins at page 92 of the Schedule. Rather than listing different restaurants, the heading “premises 11” is given with “Floater” then “GBF” (“Great Britain Floater”) beneath. This is repeated four pages later.

Under “Property insured” in the business interruption section:

Gross profit is insured on an “All Risks” basis for £67,643,191 with an uplifted sum of £90,188,666 with an indemnity period of 24 months.

Additional increased cost of working is insured on an “All Risks” basis for £514,000 with an uplifted sum of £685,316 with an indemnity period of 24 months.

Tronc (Footnote: ¹) is insured on “All Risks” basis for £5,951,627 with an uplifted sum of £7,935,304 with an indemnity period of 12 months.

The total sum insured is £98,809,286.

Among the “Cover limits” listed in the business interruption section is NDDA. The “sums insured/limits” are stated to be “100% of the sum insured or £250,000 whichever is less”.

The nature of the NDDA cover is set out in the “Business interruption section” of the Full Policy Wording. It is in the following terms:

physical damage to property at your premises or elsewhere

strikes, picketing, labour disturbances or trade disputes

the condition of or the business conducted within your premises, or any other premises owned or occupied by you

notifiable diseases as detailed in the Murder, suicide or disease cover

actions where you have been given prior notice.”

The factual background so far as concerns COVID-19 is common ground and was set out in an agreed statement of facts.

The COVID-19 pandemic reached the United Kingdom in or around early March 2020.

On 20 March 2020, the Prime Minister Boris Johnson announced “all the formations of the United Kingdom, all the devolved administrations” were “collectively telling, telling cafes, pubs, bars, restaurants to close tonight as soon as they reasonably can, and not to open tomorrow. Though to be clear, they can continue to provide take-out services.”

On 21 March 2020, the Health Protection (Coronavirus, Business Closure) (England) Regulations 2020 (“the 21 March Regulations”) were made by the Secretary of State for Health and Social Care pursuant to powers under the Public Health (Control of Disease) Act 1984 (“the 1984 Act”). Regulation 2 provided as follows:

Regulation 3 of the 21 March Regulations provided that contravention of Regulation 2 without reasonable excuse was an offence, punishable on summary conviction by a fine. Regulation 4(1) provided that a person designated by the Secretary of State may take action as necessary to enforce a closure or restriction imposed by Regulation 2.

The businesses listed in Part 1 of the Schedule included restaurants, cafes, bars and public houses. Those businesses were required to close or cease carrying on the business of selling food and drink other than for consumption off the premises under Regulation 2(1).

On 23 March 2020, the UK Government issued guidance to businesses about closures. The advice included that it would be an offence to operate in contravention of the 21 March Regulations and that businesses in breach of the regulations would be subject to prohibition notices and potentially unlimited fines. On the same day, the Prime Minister explained that “If you don’t follow the rules the police will have the powers to enforce them, including through fines and dispersing gatherings.”

On 26 March 2020, the Health Protection (Coronavirus, Restrictions) (England) Regulations 2020 (“the 26 March Regulations”) were made by the Secretary of State for Health and Social Care pursuant to powers under the 1984 Act.

The 26 March Regulations revoked most of the 21 March Regulations and introduced a more expansive regime for business closures, as explained below.

Regulation 3 of the 26 March Regulations defined the "emergency period" as follows:

Regulation 4 of the 26 March Regulations was in a similar form to Regulation 2 of the 21 March Regulations and provided as follows:

Regulation 5 introduced new closures for retail shops, holiday accommodation providers and places of worship.

Regulation 6(1) set out the general prohibition against people leaving the place where they were living “without reasonable excuse”. Regulation 6(2) provided a non-exhaustive list of reasonable excuses, including the need to obtain basic necessities, take exercise, seek medical assistance, or fulfil a legal obligation. Regulation 7 prohibited gatherings in public places of more than two people other than in limited circumstances.

Regulation 8(1) provided “relevant persons” with the power to take such action as necessary to enforce any requirements imposed by Regulations 4, 5 or 7. Under Regulation 8(3) and (4), where a relevant person considered that a person was outside the place they were living contrary to Regulation 6(1), a relevant person could direct them to return to the place they were living or remove them to the place where they were living, including by using reasonable force if necessary. “Relevant person” was defined in Regulation 8(12)(a) to include a constable, a police community support officer or a person designated by the Secretary of State.

Regulation 9(1) provided that a contravention of Regulations 4, 5, 7 or 8 without reasonable excuse was an offence and any contravention of Regulation 6 was an offence. Such offences were punishable on summary conviction by a fine (Regulation 9(4)).

Regulation 10(1) provided “authorised persons” with powers to issue fixed penalty notices. “Authorised person” was defined to include a constable, a police community support officer or a person designated by the Secretary of State.

Regulation 11 provided that the Crown Prosecution Service, and any person designated by the Secretary of State, could bring proceedings for an offence under the regulations. Regulation 12(1) provided that the regulations would expire at the end of the 6 month period beginning with 26 March 2020.

Schedule 2 to the 26 March Regulations was entitled “Businesses subject to restrictions or closure”. Part 1 referred to:

Part 2 included businesses such as cinemas, theatres and other entertainment venues.

On 4 April 2020, the Secretary of State for Health and Social Care designated local councils, including district councils, county councils and London borough councils, as “relevant persons” and “authorised persons” under the enforcement provisions in the 26 March Regulations. Specifically, local councils were empowered to take action and issue fixed penalty notices under Regulations 8 and 10 for the enforcement of Regulations 4 and 5. They were also empowered under Regulation 11 to bring proceedings for an offence under Regulations 4 and 5.

On 10 May 2020, the Prime Minister made an announcement that steps would gradually be taken to modify the measures in place.

On 4 July 2020, the 26 March Regulations were revoked and replaced with more limited restrictions in the Health Protection (Coronavirus, Restrictions) (No. 2) (England) Regulations 2020 in England (“the 4 July Regulations”). Under the 4 July Regulations, restaurants, cafes and bars were permitted to reopen.

In September 2020, it became clear that the incidence of COVID-19 was increasing. On 22 September 2020, the Prime Minister made a televised address announcing new measures. He stated that he was setting out “a package of tougher measures in England – early closing for pubs, bars; table service only; closing businesses that are not covid secure; expanding the use of face coverings, and new fines for those that fail to comply”.

On 24 September 2020, the Health Protection (Coronavirus, Restrictions) (No 2) (England) (Amendment) (No 5) Regulations 2020 (“24 September Regulations”) were made by the Secretary of State for Health and Social Care in exercise of powers conferred by the 1984 Act. Regulation 4A provided inter alia that:

Restricted businesses and services were defined in Schedule 3, Part 1 as including:

Incidence of COVID-19 continued to increase. On 31 October 2020, the Prime Minister announced a second national lockdown for the period 5 November 2020 to 2 December 2020. He stated inter alia that “Pubs, bars, restaurants must close except for takeaway and delivery services …”.

On 5 November 2020, the Health Protection (Coronavirus, Restrictions) (England) (No. 4) Regulations 2020 came into force (“5 November Regulations”). Regulation 15 required that persons responsible for carrying on a restricted business must close any premises or part of the premises in which food or drink were provided for consumption on the premises and cease providing food or drink for consumption on the premises. Regulation 16 required persons responsible for carrying on restricted businesses to cease carrying on that business or providing that service.

As under the 24 September Regulations, restricted businesses were defined to include “Restaurants, including restaurants and dining rooms in hotels or members’ clubs”, “Cafes [subject to the same limited exceptions]” and “Bars, including bars in hotels or members’ clubs”. Under Regulation 19, “relevant persons” with power to “take such action as is necessary to enforce any restrictions imposed by these Regulations” included constables, police community support officers, persons designated by a local authority for the purposes of the regulation (insofar as concerned enforcement of business restrictions and closures) and persons designated by the Secretary of State.

The closures imposed by the 5 November Regulations continued until 2 December 2020.

In this judgment, I refer to the 21 March, 26 March, 4 July, 24 September and 5 November Regulations collectively as “the Regulations”.

The Claimants appear to have intimated a claim under the Policy in the first half of 2020. On 20 November 2020, the Defendant informed the Claimants that they were not entitled to an indemnity against the loss they had suffered as a result of the COVID-19 pandemic.

Central to this case are the decisions of the Divisional Court (Footnote: ²) and the Supreme Court in the test case of Arch. While I will return to consider a number of aspects of each judgment in detail below there is much important background which requires to be stated first.

A decision was taken by the FCA to pursue a test case on behalf of the many insureds affected. The claim brought was heard on an expedited basis under the Financial List Test Case Scheme. Unusually it was heard not by a single judge of the Commercial Court, but by a two judge panel, composed of Flaux LJ (the Supervising Lord Justice of the Commercial Court) and Butcher J.

While their judgment is not entitled “Divisional Court” the court was composed as a Divisional Court would be. It is recorded in [3] of the judgment that the court was “in effect” a Divisional Court. Overtly the route to this procedure lay in the Financial Market Test Case Scheme which provides at 6.5(d):

However equally the proceedings could be constituted as a matter to be heard by a Divisional Court pursuant by CPR 3.1(2)(bb) which, further to the Civil Procedure (Amendment No. 2) Rules 2017 (SI 2017/889), provides that “except where the Rules otherwise provide, the court may - … require that any proceedings in the High Court be heard by a Divisional Court of the High Court”. This provision reflects the preservation of the rump of the practice of hearing issues of law en banc which preceded the Judicature Acts.

The reason why the decision (highly unusual in the Commercial Court) was taken to constitute a two-judge court was plainly an intention to give the judgment greater authority than would be the case if any single judge of this court were to hear it; and doubtless also to facilitate what happened – a leapfrog appeal to the Supreme Court.

In Arch, the Divisional Court therefore heard a test case to determine issues of principle in relation to policy coverage for business interruption losses arising in the context of the COVID-19 pandemic and the advice of and restrictions imposed by the UK Government in consequence.

In the course of two Commercial Court weeks the Court considered a very significant number of issues. It considered 21 different “lead” policies issued by a range of insurers; and within those policies a range of clauses, divided broadly into three categories: (i) disease clauses; (ii) “hybrid” clauses (i.e. policy terms which refer to both restrictions imposed on the premises and to the occurrence or manifestation of a notifiable disease); and (iii) clauses covering prevention of access and similar perils.

It is fair to say that the centre of gravity of the arguments before the Divisional Court (and the Supreme Court) was the argument relating to Disease Clauses. That is reflected by the fact that nearly 50 pages and over 160 paragraphs of the judgment relate to this topic. Many of the issues dealt with by the Divisional Court are not relevant to the present case. I set out below only the key clauses and passages for present purposes.

I begin by setting out the Divisional Court’s approach to certain NDDA, “action of competent authority” (“AOCA”) and prevention of access clauses. The Court generally found that such clauses did not provide cover in respect of the consequences of COVID-19. I then discuss the Court’s approach to disease clauses, which by contrast were generally found to provide effective cover.

The most common form of NDDA in the Hiscox policies considered by the Divisional Court was as follows:

Referring to (i) the use of “an incident”, which the Court considered should be understood to mean “something which happens at a particular time, at a particular place, in a particular way”, and (ii) the geographical restriction that the incident should occur within one mile, the Divisional Court held at [406] that “this clause is intended to cover local incidents, of which the paradigm examples are a bomb scare or a gas leak or a traffic accident”, the Divisional Court concluded that “[w]e consider that this is a narrow, localised cover intended to insure events or incidents which occur within the one mile radius.”

It followed from this in the Divisional Court’s view that there was no cover under the NDDA clause in respect of business interruption losses caused by the restrictions imposed by the government in response to the national pandemic.

The Divisional Court then considered three AOCA clauses in MS Amlin Underwriting Limited (“MSA”) policies. The first, MSA 1, was in the following terms:

Mr Kealey QC, for the insurers, accepted that the risk of a disease in the vicinity would qualify as a “danger” although he denied that insurers were liable on the grounds that the cover was localised and was not intended to respond to the effects of a pandemic.

The Divisional Court accepted this submission, holding at [436] that the words “following a danger or disturbance in the vicinity of the premises” demonstrated that:

At [437], the Divisional Court added:

MSA 2 was materially identical to the Hiscox NDDA. The Divisional Court accordingly found that the same reasoning applied as in respect of the Hiscox NDDA.

MSA 3 was wider than MSA 1 because it encompassed not only prevention of access but prevention of use and hindrance of both access and use. The Divisional Court found that the same considerations as in MSA 1 in respect of the wording “in the vicinity of the premises” nevertheless applied, with the result that the clause was found to provide narrow localised cover.

At [445-476], the Divisional Court considered various Royal & Sun Alliance Insurance Plc policy wordings.

RSA 2.1 included a “Prevention of Access-Public Emergency” clause setting out “What is covered” in the following terms:

Under “What is not covered”, the policy provided:

The wording in RSA 2.2 was very similar.

RSA 4 referred to

The term “Prevention of Access-Non Damage” was defined as:

“Vicinity” was defined as:

The Divisional Court held in relation to RSA 2.1 and 2.2 that, as in the case of the MSA 1 AOCA clause, the “principal difficulty” facing the FCA was [466]:

The court did not consider that “the entire country” could be “described as in the vicinity of the insured premises on the wording of this policy” ([466]). As a result, there could only be cover if the insured was able to demonstrate that it was an emergency by reason of COVID-19 in the vicinity, in the sense of the neighbourhood, of the insured premises, as opposed to the country as a whole. The court considered it “highly unlikely that that could be demonstrated in any particular case.”

At [471], the Divisional Court noted that the wording of the prevention of access clause in RSA 4 was strikingly different because it did not require that the actions or advice of the government were following or due to an emergency or danger in the vicinity of the insured premises. Instead, all that was required was that the actions or advice of the government were “in the Vicinity of the Insured Locations.” On this basis, the court considered that:

The court thus concluded that coverage was in principle available under RSA 4.

The Divisional Court also considered two AOCA clauses contained in Zurich Insurance Plc policies. They were in materially identical form for present purposes. Taking Zurich 2 by way of example, they provided:

The Divisional Court noted that this AOCA clause was very similar to the MSA 1 AOCA clause. In relation to the words “following a danger or disturbance in the vicinity of the premises”, the Divisional Court took the view at [499] that “this clearly indicates that this is narrow localised cover intended to cover dangers occurring in the locality of the insured's premises, of which the paradigm example is a bomb scare.”

The Court also referred at [500] to the fact that the word “danger” gained some colour from its juxtaposition with “disturbance”. The paradigm example of the latter would be an affray or brawl. The overall phrase contemplated “an incident specific to the locality of the premises rather than a continuing countrywide state of affairs.” Government action in response to the national pandemic accordingly could not be said to be following a danger in the vicinity of the insured premises [501].

At [307]-[336], the Divisional Court also considered an Arch prevention of access clause. The clause referred to:

The argument before the Divisional Court focused on what would constitute a prevention of access for the purposes of the clause in respect of different types of premises. It is accordingly not relevant to set out the Divisional Court’s conclusions here.

By contrast with the hybrid and prevention of access clauses, the Divisional Court generally took the view that disease clauses did not distinguish between the consequences of events occurring within the radius and those not doing so. As a result, losses suffered by reason of restrictions imposed in response to the national pandemic were covered.

One example of this is RSA 3. The “Infectious Diseases” clause in the RSA 3 policy provided as follows:

In considering the meaning of an “occurrence of a Notifiable Disease”, the Divisional Court took the view at [93] that there would be an occurrence of COVID-19 when at least one person was infected with COVID-19 within the relevant area. As to the type of causation required, the Divisional Court considered that it was necessary that one of (a) to (d) should have a causal connection to the business interruption, but not necessarily one of proximate causation.

Significantly, the Divisional Court rejected the argument made on behalf of RSA to the effect that the insured peril was the effect of a local occurrence of a Notifiable Disease. At [102]-[103], the court identified two matters which it considered fundamental to its rejection of this argument: (i) the wording of the policy - the relevant provision was not expressly or impliedly confined to cases where the interruption had resulted only from the instance(s) of a Notifiable Disease within the 25 mile radius, as opposed to other instances elsewhere; and (ii) the implications of the fact that the cover related to occurrences of a Notifiable Disease.

On this second point, the court noted that: (i) Notifiable Diseases included diseases capable of widespread dissemination in highly complicated and unpredictable patterns; and (ii) the list of diseases included some which might attract a response which was not purely local from authorities which were not local. In the court’s view, the parties must have contemplated that the authorities might take action in relation to the outbreak of a notifiable disease as a whole, and not to particular parts of an outbreak.

The disease clause in RSA 3 accordingly provided coverage. The Divisional Court applied a similar approach in respect of five other disease clauses.

There were, however, two QBE UK Limited (“QBE”) disease clauses where the court did not take this approach: QBE 2 and QBE 3. I deal with them now.

QBE 2 by Clause 3.2.4 included coverage in respect of:

QBE 3 included coverage in similar terms, save that the radius was one mile.

Based on factors such as the reference to “events” at the beginning of the clause and the reference to “the incident” in clause (h) and elsewhere, the Divisional Court concluded at [234] (in respect of QBE 2) that “there is cover only if there is business interruption as a result of the ‘event’ of the person(s) sustaining that illness within the area.”

The court reached the same conclusion in respect of QBE 3 at [237], stating that the fact that the radius was one mile made it “all the easier to accept that this is what the parties have agreed, because that reinforces the view that what is being contemplated is specific and localised events.”

The court considered at [235] that in light of this construction, “the issues as to causation largely answer themselves.” The court expanded:

Coverage was accordingly not available under the QBE 2 and 3 disease clauses.

By Order of 2 October 2020, the Divisional Court made a consequentials declaration which reflected its findings, including those that I have set out above.

The case then proceeded directly to the Supreme Court, where various appeals and cross-appeals were heard before judgment was given on 15 January 2021.

As noted by the Supreme Court at [122], the Divisional Court’s decision that certain clauses (the MSA 2 and Hiscox NDDA clauses, the MSA 1 and Zurich AOCA clauses) did not cover losses arising from the COVID-19 pandemic was not challenged on appeal. In relation to the NDDA Clauses, the only points specifically appealed were:

whether prevention of access and hybrid wordings, in certain of the policies, were triggered by actions without force of law;

the extent of the prevention or denial of access, interruption, closure, restriction or inability to use required by the prevention of access and hybrid wordings in certain policies, with the relevant terms being “ prevention ” or “ denial of access ”, “ interruption ”, “ inability to use ”, or “ closure or restrictions placed on the premises ”.

The Supreme Court did, however, deal with the coverage provided by the disease clauses. Lords Hamblen and Leggatt (with whom Lord Reed agreed) held at [95] that:

In other words, therefore, the Supreme Court considered that cover under the disease clauses was intended to be confined to the consequences of specific local cases. This was principally because ([69]), in considering the words “occurrence of a Notifiable Disease”, “[t]he 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.”

This did not mean, however, that losses arising from COVID-19 were not covered. At [161], the Supreme Court explained that this interpretation of the disease clauses meant that questions of causation became of “crucial importance”.

The majority of the Supreme Court held at [176] that each case of COVID-19 was a concurrent cause of the restrictions imposed by the government. While it “obviously could not be said that any individual case of illness resulting from Covid-19, on its own, caused the UK Government to introduce restrictions”, the government measures were “taken in response to information about all the cases of Covid-19 in the country as a whole”. All cases were therefore “equal causes of the imposition of national measures”.

Having so found, the Supreme Court considered that, contrary to the reasoning of the Divisional Court, each local case could therefore properly be said to have caused the national restrictions which interrupted business. At [191], the majority held that:

Turning then to the interpretation of the disease clauses, the Supreme Court made the following findings at [192-212]:

Matters of background knowledge to which the Divisional Court attached weight in interpreting the policy wordings were important. In view of the nature of the spread of some infectious diseases, no reasonable person would suppose that, if an outbreak of an infectious disease occurred which included cases within such a radius and was sufficiently serious to interrupt the policyholder's business, all the cases of disease would necessarily occur within the radius.

The parties could not reasonably be supposed to have intended that cases of disease outside the radius could be set up as a countervailing cause which displaced the causal impact of the disease inside the radius. This interpretation was reinforced by the fact that the relevant wordings did not confine cover to a situation where the interruption of the business had resulted only from cases of a notifiable disease within the radius, as opposed to other cases elsewhere.

The “weighing” approach contended for in the alternative (which involved aggregating all the cases of disease which fell within the scope of the policy and to ask whether those cases, taken together, had an equal or similar causal impact when compared with the aggregate impact of all the cases of disease not covered by the policy) would operate in a similarly whimsical way by setting up competition between cases occurring inside and outside the radius. The parties could not rationally have intended this.

By contrast, an interpretation which recognised the causal requirements of the policy wordings as being satisfied in circumstances where each case of disease informed a decision to impose national restrictions served to avoid irrational effects and the need for arbitrary judgments. It was also “clear and simple to apply”.

The proper interpretation of the disease clauses was accordingly that in order to show that loss from interruption of the insured business was proximately caused by one or more occurrences of illness resulting from COVID-19, it would be sufficient to prove that the interruption was a result of Government action taken in response to cases of disease which included at least one case of COVID-19 within the geographical area covered by the clause.

The Supreme Court held at [213] that the analysis in respect of disease clauses was also applicable to “those hybrid clauses which contain, as one element, an occurrence of an infectious disease within a specified distance of the insured premises.” The court noted, however, that unlike the disease clauses, hybrid and prevention of access clauses often specified more than one condition which must be satisfied, and often required certain elements to operate in a causal sequence.

The Supreme Court then examined a public authority clause in “Hiscox 1-3” which covered financial losses “resulting solely and directly from an interruption to your activities caused by … your inability to use the insured premises due to restrictions imposed by a public authority during the period of insurance following … an occurrence of any human infectious or human contagious disease, an outbreak of which must be notified to the local authority” and an Arch prevention of access clause which indemnified the policyholder in respect of reduction in turnover resulting from:

In relation to the Hiscox clause, the Supreme Court held at [243] that the clause indemnified the policyholder against the risk of all of the elements of the insured peril acting in causal combination to cause business interruption loss. The court considered that it did so “regardless of whether the loss was concurrently caused by other (uninsured but non-excluded) consequences of the COVID-19 pandemic which was the underlying or originating cause of the insured peril.”

Similarly, the Supreme Court found that the Arch clause provided cover for losses arising out of the COVID-19 pandemic, rejecting the notion that the indemnity was confined to loss which would not have occurred but for the operation of the insured peril.

At [250], the Supreme Court acknowledged that while it was “unnecessary to address other hybrid and prevention of access clauses”, “[i]n principle … a similar analysis must apply to those clauses as to the clauses which we have specifically addressed.”

The matters decided by the Supreme Court were reflected in an Order dated 13 July 2021. The Supreme Court ordered that the declarations appealed be varied as set out in an annex to the Supreme Court’s Order. The Supreme Court only ordered variation in relation to matters that had been appealed.

The Irish High Court gave judgment in Brushfield Limited (t/a The Clarence Hotel) v AXA Insurance Designated Activity Company & Anr [2021] IEHC 263 on 19 April 2021. The Court was considering inter alia an NDDA clause in the following terms (materially identical to that which arises for consideration in this case):

The murder, suicide or disease (“MSDE”) clause contained a list of specified human infectious or human contagious diseases. COVID-19, SARS, severe acute respiratory syndrome coronavirus 2 or SARS-CoV-2 were not included in the list of specified human infectious or human contagious diseases in the MSDE Clause.

The court held at [187] that:

The court then addressed the question of whether COVID-19 constituted such a “danger or disturbance”. At [190], the Court said:

In reaching this conclusion, the court relied on a number of factors (set out at [190(a)-(f)]). These included:

The fact that the clause was concerned not with dangers or disturbances simpliciter, but dangers or disturbances at the insured’s premises or within one mile of the premises. The court referred in this regard to the Divisional Court’s view in FCA v Arch that reference to the “vicinity” had a local connotation and supported the conclusion that the intention of the provision was to address something arising at local level.

The juxtaposition of the word “danger” with “disturbance” had a “very obvious local connotation”.

Reference to “actions” by the police or statutory body should be read as “intended to extend to measures which do not have the force of law such as those which may have to be taken urgently to address an immediate danger”. Such danger would typically arise at local level.

The court could not “see how it could plausibly be contended that the measures taken at a national level by the Government or the Minister for Health could be said to have been proximately caused by a risk of COVID-19 within a one-mile radius of the hotel.”

The exclusions to the clause were concerned with situations where “advance notice directed to the plaintiff is given of proposed actions that are likely to hinder or restrict access.” This supported the conclusion that the clause was local in nature.

The court accordingly concluded that the clause did not provide cover in respect of losses caused by restrictions imposed in response to the COVID-19 pandemic.

It is perhaps unusual to cite an arbitration award; but these are unusual times -and the China Taiping Award is an unusual award.

In June 2021 a group of 183 hospitality policyholders and the China Taiping Insurance (UK) Co Ltd agreed an ad hoc arbitration agreement, to submit certain coverage issues to arbitration by Lord Mance, the former Deputy President of the Supreme Court, and (as is well known) a noted authority on insurance law, as sole arbitrator.

On 10 September 2021, Lord Mance issued an award in The Policyholders Specified in Schedule 1 to the Arbitration Agreement v China Taiping Insurance (UK) Co Ltd (“the China Taiping arbitration”).

In those arbitration proceedings, coverage was considered under two limbs of a Denial of Access clause which responded to:

There were three key disputed issues. First, whether the existence of Notifiable Disease cover elsewhere in the policy (which did not extend to Covid-19) negated the possibility of the Denial of Access wording responding to the pandemic. Secondly, whether the requirement for an “emergency in the vicinity of the premises” in limb (c) meant that the clause could only respond to narrow, localised events, rather than national ones, per the Divisional Court decision in the test case. Thirdly, whether the UK Government was a “competent local authority” within the meaning of the clause.

On the first issue, Lord Mance found in favour of the policyholders. The existence of the express notifiable disease cover elsewhere in the policy did not limit the cover under the prevention of access extension. It was common for the coverage provided by various insuring clauses and extensions to overlap, and if insurers intended to exclude diseases from the scope of the prevention of access clause, they should have used clear language to do so.

On the third issue, Lord Mance agreed with insurers that the UK Government was not a “competent local authority” within the meaning of the clause, meaning that there could be no coverage under limbs (b) or (c) of the Denial of Access extension for losses caused by closures and other restrictions imposed by the UK Government in response to the Covid-19 pandemic. This issue was ultimately therefore fatal to the policyholders’ claim, which failed at the last hurdle.

Causation was therefore not strictly speaking live. However Lord Mance, doubtless appreciating the wide interest in the point, dealt with the second issue in any event. The basis on which he did so is worth noting:

On causation, Lord Mance noted at [65-66] that while the causation required under an insurance policy is a matter of construction,

Lord Mance concluded that he was not bound by the Divisional Court’s reasoning in the FCA v Arch case.

As set out by the Supreme Court of New Zealand in New Zealand Fire Service Commission v Insurance Brokers Association of New Zealand Inc [2015] NZSC 59, the term “composite policy” derives from the judgment of Sir Wilfrid Greene MR in General Accident Fire and Life Assurance Corp Ltd v Midland Bank Ltd [1940] KB 388 (CA). It connoted in Sir Wilfrid Greene MR’s view a policy which compromises in one document “the interests of a number of persons whose connection with the subject matter of the insurance makes it natural and reasonable that the whole matter should be dealt with in one policy.”

As set out in MacGillivray on Insurance Law (14^th ed.) at 1-202:

In New Hampshire Insurance Co Ltd v MGN Ltd [1997] 1 LRLR 24, one of the issues was whether the cover provided was single, joint or composite (meaning on the court’s definition that there was more than one assured and each assured was insured separately) in circumstances where several different companies within the Maxwell group were named in the policy. In addressing this issue, the Court of Appeal noted that the term “insured” covered two aspects of the person described: (i) the person with whom the insurer contracts; (ii) the person who is interested in the property or other event covered by the insurance. The court concluded that:

The Parties agree that access to the premises of each of the Claimants was restricted or hindered as a result of government regulations imposed in response to the COVID-19 pandemic in the following manner:

The Claimants’ premises were forced to close from 20 March 2020 until 4 July 2020 (“March 2020 Closure”).

An enforced closing time of 10.00 pm was introduced on 24 September 2020 (“September 2020 Restriction”).

The Claimants’ premises were forced to close again from 5 November 2020 to 2 December 2020 (“November 2020 Closure”) (collectively, “the Restrictions”).

The parties differ on the Coverage Issue and the Quantum Issue. I deal with their submissions on those issues below. Before doing so, I address a preliminary issue between the parties concerning whether certain of the Claimants have any arguable cause of action.

Axa submits that five of the Claimants do not have any arguable cause of action against the Defendant. These Claimants are: Corbin & King Limited (“C1”); The Wolseley Restaurant Property Limited (“C4"); The Delaunay Restaurant Property Limited (“C6”); Brasserie Zedel Property Limited (“C8”), and; The Bellanger Restaurant Limited (“C11”).

This is because:

The Claimants accept in their Reply that no claim was being made by C1.

The Claimants state that C4, C6 and C8 are the property companies of the various operating companies. Yet a table of property interests provided by the Claimants shows that the leaseholder of C8’s premises is in fact Corbin & King Restaurant Group Limited (“C2”).

A letter from the Claimants’ solicitors dated 6 August 2021 states that: the Third Claimant, not C4, suffered loss in respect of part of 157-160 Piccadilly and Arlington Street (The Wolseley Restaurant); the Fifth Claimant, not C6, suffered loss in respect of 51 Aldwych (The Delaunay); the ninth Claimant, not C8, suffered loss in respect of Sherwood Street Restaurant, Sherwood Street.

The Claimants admit in their Reply that the Bellanger Restaurant was closed in August 2019 for reasons unconnected with COVID-19 and was reopened by C2 in August 2020. The Claimants have accepted in correspondence that the loss in respect of the Bellanger Restaurant was suffered by C2, not C11.

Ultimately the agreed position on this was that in the event I were to conclude that the Claimants' claim succeeds there should be no declarations made in respect of those claimants and that on that hypothesis at the quantum/adjustment stage, if the Fourth, Sixth and Eighth Claimants otherwise have an admissible claim under the Policy Axa will not seek to rely on their absence from the proceedings.

In relation to the Coverage Issue, the issue is:

Whether, as the Claimants contend, they are entitled to an indemnity under the NDDA clause provided they can demonstrate:

That there were cases or the threat of cases of COVID-19 at or within a one-mile radius of each of the Claimants’ premises; and

Such cases or threatened cases, combined with actual or threatened cases elsewhere in the UK, were an effective cause of the passing of the Regulations which led to the restriction of access to each of the Claimants’ premises.

Or, alternatively, whether, as Axa contends:

The NDDA clause only provides a narrow, localised form of cover in respect of “a danger or disturbance” specific to the locality of the Claimants’ premises occurring at the Claimants’ premises or within a one-mile radius as opposed to a nationwide state of affairs; and

The Claimants are only entitled to an indemnity under the clause if they can demonstrate that it was the presence or the risk of Covid-19 at the Claimants’ premises or within a one-mile radius, as opposed to the country as a whole, which led to the Regulations.

If I accept Axa's submissions on the Coverage Issue, the Claimants will not seek to demonstrate that the presence or risk of COVID-19 at the premises or within a one-mile radius thereof led to the Restrictions. Similarly, if I accept the Claimants’ submissions on the issue, Axa accepts that the Claimants have shown that there was at least one case of COVID-19 within a one-mile radius of each of the Claimants’ premises prior to 20 March 2020.

The Claimants submit that:

COVID-19 is a “danger” to life and health within the meaning of the NDDA.

There must be a danger within the one mile radius of the premises.

At all material times from at least early March 2020, there were actual or threatened cases of COVID-19 which constituted such “danger” at the Claimants’ premises insured by the Policy or within a one-mile radius of each of them.

The passing of the Regulations which led to the restriction of access was caused by the danger which was constituted by every single actual or threatened case of COVID-19 at each of the Claimants’ premises and/or within one mile of them, as well as by other actual or threatened cases of COVID-19 elsewhere in the UK. Each actual or threatened case was of equal potency and a separate and effective cause.

The Regulations which mandated the closure of the Claimants’ premises were passed in response to the “danger” posed by COVID-19.

It is not a requirement of the Policy that the “danger” should be exclusively present within that one-mile area.

The NDDA thus provides coverage for losses arising from the Restrictions.

At the centre of the submissions is the contention that the Divisional Court judgment in FCA v Arch that NDDA clauses such as the present do not provide an indemnity in the circumstances set out in (1) to (4) above cannot stand in the light of the Supreme Court’s analysis of the doctrine of causation.

The Claimants submit that the Divisional Court was wrong in that:

It decided that, because business interruption losses were caused by restrictions introduced to avert the dangers arising from the pandemic in the UK as a whole rather than in any particular localised area, coverage could only be available if, on their proper construction, disease or NDDA clauses provided cover against the effects of the pandemic in the whole of the UK.

It mistakenly found that those clauses which provided cover in respect of business interruption losses caused by “dangers” or “emergencies” limited to a radius of 1 or 25 miles of the premises, did not indemnify an assured against business interruption losses caused by regulations introduced to deal with the national pandemic.

Having determined the issues before it on construction, the Divisional Court wrongly deemed the issue of causation to be irrelevant.

Because the Supreme Court decided that disease and NDDA clauses did, indeed, only provide cover limited to the effects of “dangers” or “emergencies” within a radius of 1 or 25 miles of the premises, an analysis of causation was essential.

On the basis of the proper consideration of the issue of causation, “dangers” or “emergencies”, occurring locally were a concurrent cause of the closures and restrictions, and the consequent business interruption losses, along with all the other infectious diseases, “dangers” or “emergencies” elsewhere in the UK.

The Claimants refer in support of their arguments to the reasoning of Lord Mance in his award in the China Taiping arbitration. They draw particular attention to paragraph 67 of the Award and say that Lord Mance’s views are “compelling, cogent and should be given substantial weight”.

The Defendant submits that the NDDA clause covers transient, localised dangerous incidents or disturbances in response to which action is taken at a local level to prevent access to the insured’s premises. The Defendant suggests paradigm examples to be unexploded bombs, structures at risk of collapse or an affray. The presence and/or risk of further cases of a disease that form part of an ongoing national and international pandemic are in the Defendant’s submission “as far removed from these paradigm cases as it is possible to be.” In “no reasonable sense” could the state of affairs brought about by the COVID-19 pandemic be described as a “danger” at the Claimants’ premises or within one mile of them.

The Defendant also relies in support of this interpretation on the reasoning of the Divisional Court in relation to NDDA and AOCA clauses in Arch. In particular, the Defendant relies on the Divisional Court’s conclusion that no cover was available under the MS Amlin and Zurich AOCA Clauses. The Defendant argues that those clauses were in all material respects the same as the NDDA cover under the Policy.

The Defendant says that the Divisional Court’s reasoning was reflected in the consequentials declaration made by Order of 2 October 2020, which passages were subsequently incorporated into the Order of the Supreme Court dated 13 July 2021.

The Defendant submits that the Divisional Court’s decision in Arch binds me unless I can be satisfied that it cannot stand with the decision of the Supreme Court. The Defendant says that its decision was, or was akin to, a Divisional Court as provided for by section 66 of the Senior Courts Act 1981 and should be treated as such.

In response to the suggestion that the Supreme Court’s reasoning in respect of the QBE2 and QBE3 disease clauses undermined the Divisional Court’s analysis of the MSA and Zurich AOCA Clauses, the Defendant submits:

The Divisional Court’s reasons for holding that the MSA and Zurich AOCA Clauses did not provide coverage were not the same as those that it relied upon in respect of the QBE2 and QBE3 disease clauses. The Supreme Court’s approach to causation is accordingly not determinative of the coverage provided by clauses in the latter group.

The FCA did not appeal the Divisional Court’s clear conclusion that the events that led to the imposition of Regulations by the government were not a “a danger or disturbance in the vicinity of the premises” or “incident” or “emergency in the vicinity”.

The cover provided by the NDDA clause in this case is qualitatively different from the cover provided by a disease clause. To suggest otherwise is to commit a category mistake.

The Claimants are mistaken to place reliance on what the Supreme Court said about the causal requirements of what were classified as “hybrid” clauses and the Arch “prevention of access” clause.

It is wrong to suggest that the Supreme Court stated at [250] that its analysis of causation in relation to disease clauses requiring the occurrence of disease within a specified radius could be applied simply to prevention of access clauses such as the NDDA cover.

The Defendant submits that the Divisional Court was accordingly right to conclude that the MSA 1 AOCA and Zurich AOCA Clauses required a conventional “but for” approach to causation, and the same approach is appropriate in relation to the Defendant’s NDDA cover in the present case.

In relation to Lord Mance's decision in the China Taiping arbitration, the Defendant seeks to emphasise that Lord Mance was dealing with a clause differently phrased to the NDDA cover, and that his comments were obiter and expressly caveated.

The Defendant also points to the decision of the Irish High Court in Brushfield which was handed down after the Supreme Court’s decision in Arch which it says supports its interpretation.

The first issue to consider is probably the extent to which I am free to diverge from the decision of the Divisional Court, and the approach which I should take to it.

In the end there was no real disagreement about this point. The position is as set out in Huddersfield Police Authority v Watson [1947] KB 842, per Goddard CJ at 848:

I am therefore bound by the decision of the Divisional Court – but only so far as (i) the point in question was one which was argued and decided in the Divisional Court and (ii) the Divisional Court’s analysis is not undermined by the decision of the Supreme Court. On the question of being bound, I therefore agree with the analysis of Lord Mance at [53-4] of the China Taiping Award:

The first layer of the argument is that of construction. If the Defendant is right about construction and/or I am bound by the Divisional Court and the clause provides cover only in relation to localised risks (as opposed to risks which manifest in whole or in part locally) then the issue of causation becomes irrelevant.

This was the approach taken by the Divisional Court to the NDDA and AOCA clauses before it. At [467] of the judgment it concluded that having decided that the cover was localised, it could not be demonstrated that the conditions for cover were met and so “the wider issues of causation and counterfactuals which arise in other contexts do not arise”.

The Divisional Court’s decision of course related to different clauses from the clause which I have to consider. I have therefore considered whether (i) the clauses which the Divisional Court had to consider and (ii) the arguments addressed to it were so similar to the clause and arguments which are before me that it would effectively be illogical to come to a different conclusion. As regards the similarities in the wording, this appears to be the conclusion which Lord Mance would have reached absent the Supreme Court’s decision because he said at the end of [54] in the China Taiping Award: “On the basis that construction is a matter of law, and there is no discernible difference in wording or context, [the judgment of the Divisional Court] may even, on the face of it, bind me.”

However in this case, having considered the clauses carefully, I do not consider that the very considerable similarities which do exist between the clauses considered by the Divisional Court and this clause are so strong that it effectively prejudges my own consideration of the Axa NDDA clause.

The clauses which the Divisional Court considered in this context were:

Arch prevention of access;

Hiscox NDDA;

MSA 1-3;

RSA 2.1, 2.2 and 4;

Zurich.

In particular the textual differences between the clauses and the Axa clause can be tabulated thus:

While I would tend to agree with Lord Mance that the difference between vicinity and a one mile radius wording might not be enough to draw a distinction between two otherwise identical wordings, there are greater disparities here. "Emergency likely to endanger life in the vicinity" may mean something different to a "danger within a mile". The word “Incident” also is capable of lending a very distinct colour to a wording – as the telling argument recorded at [398] of the Divisional Court's judgment makes clear. Its significance is underpinned by the Divisional Court's conclusion at [404] that “In our judgment, the FCA’s entire case on the NDDA founders on the requirement for “an incident”.”

I would also tend to the view that the precise wording of the authority or the length of the franchise period might well have an impact. For example "incident" teamed with an authority wording led off by "local authority, police, emergency services …" and a short franchise period would certainly have a real sense of pointing to the paradigm situation. Then there are other factors which can feed in to the overall construction of the clause - such as the disease exclusion, to which I will come below. It is unlikely that that is the only other factor which could impact.

I would if necessary conclude that there are sufficient differences either in overall content or terminology or both between those clauses and the Axa clause to permit of my approaching this clause from first principles.

But there is a second and important distinction which squarely engages Lord Mance’s rider of “the relevant point having been squarely argued and decided”. That consideration of the clauses in the Arch case was based on a line of argument which appears to have been subtly but significantly different from that which was before me. Before the Divisional Court it was argued that radius and vicinity clauses offered a “narrow, localised” form of cover – and that argument was accepted by the Court. But one must pay close attention to what is and was meant by “narrow, localised” in the context of the argument advanced.

In Arch one might say that the construction issue was the same as the one I face, in that the submissions were addressed to the question of whether the facts triggered the clause on its true construction, and hence (broadly) whether COVID-19 was covered by the clause on its true construction. Thus in respect of (for example) the Hiscox NDDA the submissions might be said to be that COVID-19 is not an "incident" within the meaning of the clause [398].

However that ignores the building blocks of the argument. The argument was about the pandemic as a risk. It was not focussed (as it is before me) on whether disease (in individual cases) can be covered. Before me the case advanced is that the cover is narrow and localised in the sense that (i) only a danger of an essentially localised kind can trigger cover and (ii) that means - on Axa's primary case - that disease is not covered at all.

The Divisional Court faced a different case for insurers which focused on the concept of narrow localised cover by reference to such terms as "vicinity" (and which regardless of precise ambit, led to a “but for” causation problem).

The difference is perhaps best seen in the different cases made for the insureds:

Before me Mr Gruder QC contends that on the true construction of the clause COVID is a danger and his clients would be entitled to an indemnity under the clause so long as they can demonstrate that it was the risk of COVID at the Premises or within a one mile radius of the Premises which led to the Regulations, irrespective of the (inter)national nature of the pandemic. A single case within the radius suffices.

Before the Divisional Court the essence of the FCA's argument was that the clauses were engaged on the basis that although COVID was everywhere, because it was everywhere it was also local, hence it was in the vicinity or radius. Essentially the danger or emergency was "the pandemic":

So for Hiscox: “the entire country can be described as in the vicinity of the insured premises” ([436]).

For the MSA wording at [424]: "Mr Edelman QC submitted that … the risk of contracting the disease was a danger everywhere in the UK and thus in the vicinity of any UK premises." matched with the finding at [436]: "we do not consider that the entire country can be described as in the vicinity of the insured premises, from which it follows that, …, the government action in imposing the Regulations in response to the national pandemic cannot be said to be following a danger in the vicinity, … of the insured premises."

For RSA: "the emergency was the pandemic, … and, since the pandemic was everywhere in the UK, occurring in all areas, it was “in the vicinity of the Premises” for the purposes of this provision.". Correlative finding: " we do not consider that the entire country can be described as in the vicinity of the insured premises on the wording of this policy."

For Zurich: "Since the danger of the disease was everywhere in the UK, the whole country was in the vicinity of the insured premises, because the danger everywhere else could reasonably be expected to impact on the insured premises." Correlative finding: "the overall phrase “a danger or disturbance in the vicinity of the Premises” contemplates an incident specific to the locality of the premises rather than a continuing countrywide state of affairs."

The Divisional Court thus faced a dichotomy of:

For policyholders: [covid is everywhere]+[proper construction is: everywhere triggers vicinity or radius]=[but for causation satisfied]

For insurers: [covid is everywhere]+[everywhere cannot sensibly mean vicinity]=[never get to causation].

That dichotomy made perfect sense in circumstances where, as the Divisional Court said:

That followed of course from the fact that at that point it was assumed that the Regulations could only be viewed as having been caused by the pandemic and not by individual cases. In effect therefore one can see from the Court's explanation of the arguments that in the Divisional Court the assumption by all involved that “but for” causation was applicable meant that the weight of the FCA construction arguments in favour of cover logically had to fall within parameters which would then be able to produce a “but for” causation result.

It was therefore never argued (at least as a primary case (Footnote: ³)) for the policyholders, as it has been before me, that the clauses in question meant that there was narrow localised cover which could nonetheless extend to a single case of COVID-19 in the vicinity. The policyholders' argument in relation to NDDA clauses was “all in” because any other approach was apparently doomed to fail.

Conversely for insurers before the Divisional Court there was no focus on the precise line drawn here. However it was at least tacitly conceded before the Divisional Court that disease could be a danger in the vicinity for the purposes of the relevant policy [430], but that acceptance was closely followed by a reminder of the causation issue:

I therefore conclude that (quite aside from the specific differences in the wordings) the issue engaged here was not squarely raised and decided. It is also not the case that (as the Claimants initially argued) the Divisional Court erred or that the Divisional Court’s analysis is undermined by the decision of the Supreme Court; but rather that the decision of the Supreme Court has moved the goalposts and the argument which has emerged is materially different.

This is not a question of equating the decision of construction as to the disease clauses with that in relation to NDDA clauses (an approach which Axa stigmatised as committing a category mistake). It is fair to say that the terminology used in relation to the two types of clause is not identical: QBE 2 and 3 and hence disease are analysed in terms of "specific (relatively) local" where NDDA clauses are seen as "narrow, localised". However the point is that either approach would prior to the Supreme Court's decision create a "but for" causation problem. The Supreme Court's approach to causation in relation to disease opened up the field for a different iteration of the construction argument, which approximates to the one adopted by the Supreme Court on disease clauses.

I do therefore conclude that I am not bound by the Divisional Court's conclusions on construction. However the analysis of the Divisional Court in relation to those similar but different clauses - and albeit predicated on a rather different argument - will plainly have considerable relevance to the issues inherent in the argument.

Firstly I address the overall structure of the argument. Here, as I have noted, the outline of the argument has developed somewhat – almost certainly in the light of the judgment of the Supreme Court as well as in the way that legal argument inevitably develops, like water, to find its way round an obstacle. The dichotomy between local (with but for causation problems) and vicinity equating to national is no longer the boundary of the argument.

The Claimants here accept the need for local manifestation inherent in the radius, but they draw the line at an exclusively local risk. They submit that on the true construction of the clause it would be wrong to say that danger can only be local.

It seems to me that the Claimants are right about this. The difficulty which presents itself is the relationship between what has been described as the paradigm coverage under an NDDA/AOCA clause and the construction of the particular NDDA/AOCA clause; and in particular to what extent that paradigm drives or informs the approach to contractual construction.

The paradigm is set out in the Divisional Court's judgment which concluded in relation to the Hiscox NDDA clause that “this clause is intended to cover local incidents, of which the paradigm examples are a bomb scare or a gas leak or a traffic accident”, also noting at [489]:

On the face of it, this is a promising start for Axa. However the more one looks at the point, the more problematic Axa's argument becomes.

The first problem is that the clause is couched in general language. While there is no reference to generalised danger, but rather to “a danger” that does not tell one very much about the danger itself. It may most obviously connote “a transient incident posing a risk to health and safety”; but it does not limit itself to that. Objectively therefore the parties can be taken to have intended coverage to extend to generic circumstances which fulfil the requirements of the generic words. Similarly with the reference to “at the premises” and the radius; there is plainly a local link, but that is not in dispute. It does not in my judgment indicate that the clause must be concerned with something of very local significance.

The second problem is that as a matter of law we do not construe contractual provisions by reference to such paradigms or such evidence of historical derivation. Both parties were in agreement as to the principles of construction. Both also emphasised the correctness in this context of the following passage in the Supreme Court's judgment:

Both parties also broadly accepted the accuracy of what might be described as “the Mance Variation” at [18] of the China Taiping Award:

But on this basis, we approach the Policy on its words, as if we are a small businessman, albeit with a broker to assist us. While factual matrix was mentioned in passing by Mr Christie QC for Axa, none has been pleaded in this case, as it should be (the requirement in the Commercial Court Guide is of long standing, dating back to the incumbency of Flaux J as Judge in Charge of the Commercial Court). Nor in reality was any really prayed in aid: Axa suggested the paradigm might be factual matrix “if those were matters which were available to the parties or could reasonably have been available to the parties”. He did not suggest that “the paradigm” or other factual matters were. This difficulty infects much of Axa's argument – for example the contention that "the correct question you have to ask is what kind of cover the clause provides".

On the correct conventional basis what the clause discloses is this. Firstly there is risk definitional wording which although not the most obvious fit, is linguistically apt to cover disease as one among the dangers covered. “Danger” is a word which can take on shades of meaning, depending on the context in which it is used. Instinctively (see the argument at DC [405] partnering emergency with danger) one can understand the association with the paradigm situations; but it is (as various insurers accepted in argument in Arch: see DC [310] [360] [430]) hard to resist the submission that a disease can be a danger – not least when the language of illness and infection sometimes uses that very word. If "danger" can cover disease, it becomes, if not the insured peril (as it is under the disease clause) but a potential component of an insured peril.

Axa did submit (fairly faintly) that danger paired with the “at/within” one mile radius meant that the kinds of risks which might straddle a radius were excluded; because it was implied by these words that if “a danger” straddles the notional one mile radius boundary, it is not covered because it is not within the radius, but, rather, both within and outside it. This seems to me to be a hopelessly artificial approach; not least when it is perfectly easy to posit examples of more paradigm dangers or disturbances which could equally well be occurring outside the radius as well as within it. It is also an approach which is at odds with the approach of the Supreme Court.

Secondly there is no locality limitation other than the radius. Danger is not itself paired with local. Nor is any relevant authority identified as local (as does happen in, for example, the Arch wording, Zurich, Taiping (c) and MSA1). While Axa contends that the wording for the authority tends to suggest locality (police coming first), it cannot escape from the fact that the wording “any other statutory body” is manifestly wide enough to encompass central government. This was accepted in Axa’s pleading and its attempts to move away from this position did not gain traction, particularly in the light of the Supreme Court’s declarations paragraph 9 which makes plain that the UK Government is a statutory authority for the purposes of clauses of this nature.

Thirdly there seems no basis for Axa's submission that a danger in this context is transient, in circumstances where, whatever the clause does cover, cover does not bite until the danger has existed for more than 2 hours (the Franchise Period); and the indemnity period is up to 12 weeks. While it is right not to conflate the indemnity period with the currency of the danger – because, for example, a hotel afflicted by a transient plague of rats as the Pied Piper passed by might still suffer business interruption from the nervous or squeamish weeks later – the 12 weeks is not without some significance. It shows that it is contemplated that some at least of the “dangers” covered by the clause can have effects which last more than 12 weeks (if not there would be no point in the cut off). Mr Gruder posited the effects of a tactical nuclear device, with a short active danger period but a long “tail” in terms of denial of access (though, as Mr Christie noted, that ignores the effect of other exclusions, such as the Radioactive Contamination exclusion).

Finally - and to my mind significantly - there is the position as regards the disease exclusion. This clause explicitly excludes cover for NDDA caused by the diseases covered by the MSDE clause. This is a point which was not to the fore before the Divisional Court or hence before the Supreme Court (Footnote: ⁴). But here it has taken considerable prominence in the submissions for the Claimants; and it poses important questions.

In particular if one adopts the persona of the reasonable SME owner (advised by his broker) and looks at the wording in this case one sees, in effect:

In other words the clause says that some diseases are excluded – with the logical correlate that not all are. The natural reading is that diseases not on the excluded list can be covered – if the other conditions are met (which, of course, in many cases they will not be). Given the different limits and structure of the MSDE cover and the NDDA cover there is no reason why the two should be mutually exclusive – a conclusion also reached by Lord Mance in the China Taiping Award.

It follows that I emphatically do not agree with the submission for Axa that:

Testing this point practically, if one were to go to look at the Policy with a question: “If I am denied access to my premises twice by a relevant authority, once because of a case of leprosy amongst my staff, and once because of a case of Ebola, would I have cover?”, the answer which would come would be: “Leprosy – cover under MSDE clause but excluded under NDDA. Under your NDDA cover you should be OK on Ebola: Danger, yes – denial of access, yes –and it’s not excluded because it is not in the MSDE clause…”.

This is effectively the answer to one of Axa’s other points – that the fact that the clause refers to “the unlawful occupation of your premises by third parties” or excludes “3 the condition of or the business conducted within your premises or other premises owned or occupied by you” and “5 actions where you have been given prior notice” – emphasises the locality of cover. This is an impermissible textual reading down, exactly contrary to the process to be expected of the SME reader.

True it may be that in reality, if one were to carefully unpick the drafting history, the exclusion may be the result of over-precautionary drafting. That was effectively the submission of Axa. But the reasonable reader, the SME owner (and probably the high street/retail broker) is not privy to the drafting process and does not know this. Further it is not as if the inclusion of the cover is nonsensical; there is a perfectly sensible rationale for a scheme which does give partial disease cover through the NDDA clause: business interruption caused by the MSDE diseases benefit from more generous cover under that specific coverage, so NDDA cover is not needed. But the MSDE list is to some extent arbitrary; there will be diseases which could lead to business interruption which are not on this list (such as novel epidemic flus or, indeed, COVID). Covering the spectrum of disease only on less generous terms and subject to the extra hurdles inherent in the NDDA wording might well appear to be a sensible, balanced way of achieving non-arbitrary disease coverage.

There are obviously points to be made against this analysis. In the skeleton for Axa, Mr Harris (to whom Mr Christie gave the appropriate author credit) put it this way: the Claimants' approach to the danger point “is rather like referring to a shoreline as a trillion grains of sand, when any reasonable person would call [it] a beach”. However that is an approach now firmly justified by the Supreme Court in the context of disease clauses, and while that does not dictate the conclusion in this context, there appears to be no reason why the same approach should not be applicable, once one reaches the point of concluding that disease is capable of triggering the clause. It is not a case of assuming a substantive equivalence which does not exist, as Axa would contend; while the cover provided by the disease clause may in one sense be qualitatively different from that under the NDDA clause, where you have disease being capable of triggering the NDDA clause there is a logical crossover for the approach to that component. And in truth Axa's argument, which was (primarily) that disease was not covered tacitly acknowledges that reality.

Further while Axa criticises the Claimants’ approach as “atomistic”, their own is effectively driven by conflating the construction question with the causation question – and in particular with a discomfort with where the Supreme Court's approach to causation goes. The approach of Axa is not a conventional approach to construction. At the end of the day the approach advocated for by the Claimants – to view COVID-19 as a danger, with manifestations within the radius, is more coherent with a straightforward reading of the clause as the hypothetical SME user would do.

It is then appropriate to consider whether other matters compel a different conclusion. Axa suggests that two things do: (i) the decision of the Divisional Court in Arch, which it says concluded that materially identical clauses offered by MS Amlin and Zurich did not provide policyholders with cover and (ii) the decision of the Irish High Court in Brushfield Limited (t/a The Clarence Hotel) v AXA Insurance Designated Activity Company & Anr [2021] IEHC 263.

Dealing first with the decisions on MSA and Zurich NDDA clauses, as I have already indicated in passing, I do not accept that these clauses were materially identical. While I would tend to agree with Lord Mance (at [65] in relation to a different part of the argument) that the distinction of “vicinity” is not enough were all else equal, all else is not equal. Neither clause had the extra dimension generated by the disease exclusion, and the possibility of disease in the vicinity/radius being demonstrated to be covered by this logical process was not in play. There are also other textual differences which together create a different picture. Both are phrased in terms of vicinity and:

MSA: the authority is phrased police/local/civil as opposed to police/any other statutory;

Zurich: the authority is phrased local/civil/military; Franchise 3 hours/12 months.

In my judgment these differences taken together with the moved goalposts which have enabled a differently iterated question mean that the conclusions on these two clauses do not compel a different result to that which the earlier analysis indicates.

As for the argument that there is anything in the declarations that were granted, this was sensibly not really pressed before me. It is quite plain from a reading of the declarations that they relate to the particular clauses before the court. And while they are not phrased as depending upon the approach to causation it is, as I have concluded above, fairly apparent that the scope of the argument was driven by that boundary issue.

I then turn to Brushfield, which is particularly interesting given that: (i) it concerns this exact clause (Axa having been the defendant there also) (ii) it post-dates the Supreme Court's judgment and (iii) it was (as Mr Christie demonstrated in oral argument) decided by a judge who in the context of another case had received extensive submissions on the implications of the Supreme Court's judgment.

At [171]–[191] of the judgment McDonald J dealt head on with the question: “Does COVID-19 constitute a ‘danger or disturbance’ within the meaning of the denial of access clause?”. However again the argument was not entirely the same as that which was deployed before me. In particular the disease clause was deployed very differently – in that case Axa rather ambitiously argued that despite not being listed in the MSDE clause, COVID-19 was nonetheless covered by it, and hence was an excluded peril. The judge had little difficulty dealing with this argument.

At [187] (before going on to conclude that the language of the clause was clear) the judge said:

However he went on to conclude that the reference to “danger or disturbance” in the denial of access clause was not intended to extend to a pandemic which has nationwide effects. As I do differ from this view, I should reproduce his reasons in full:

In essence I regard this reasoning as being overly concerned with the paradigm and inferred subjective intention and not enough with the wording; and also as having to some extent conflated the reasoning applicable to the causation stage of the argument with the pure construction argument. In particular:

The decision has to encompass many issues – including the disease exclusion argument to which I have alluded and another argument based on the inclusion of encephalitis in the closed disease list;

The starting point is that the Divisional Court approach is taken, without noting the effect of the dichotomy which drove that analysis. It appears that the judge may not have had it put to him that in this context something may be a danger locally while other local dangers are occurring in other localities or even nationally; or that the Supreme Court’s analysis called into question the assumptions upon which the Divisional Court’s reasoning was based. Further of course he did not have the benefit of the China Taiping Award;

The other reasons then have the appearance of being added to bolster that conclusion;

There was a good deal of evidence called which seems to have been deployed as factual matrix evidence, though there is no examination of whether it properly qualifies as factual matrix evidence bearing in mind the test for construction which I have outlined above;

The argument at (b) that the danger/disturbance juxtaposition drives construction is not a good basis for construction. One word can have a local connotation and the other not. At the same time the judge does not note the absence of references to locality aside from radius. In addition this begs the local (single case) approach. The last line of (b) overtly suggests reasoning driven by the paradigm;

The approach to the word “actions” in (c) again appears to be paradigm driven (for example in the reference to "typically") and to ignore the facts that "actions" on its face (to the reasonable SME reader) can include statutory measures. It also omits to consider the other side of the equation: that the authority wording is very broad: “any other statutory body”;

Sub-paragraph (d) strays into the causation issue, for example in saying: “I cannot see how it could plausibly be contended that the measures taken at a national level by the Government or the Minister for Health could be said to have been proximately caused by a risk of COVID-19 within a one-mile radius of the hotel”. Such considerations should be discounted in the context of construction;

In (e) the judge considers (as I have done) the relevance of the exclusions, but really focusses only on the notice point. The fact that in some cases notice may be given does not impel any conclusion about the range of actions which may be covered; it only illustrates one kind of case which will be caught by the exclusion. This is essentially the point I have covered at paragraph 190 above;

Paragraph (f) explicitly invokes the paradigm, which is dubious given that there is no clear finding that the nature of the origin of the clause was reasonably available and at [176] the judge indicates that he would not put too much emphasis on the historical origin of the clause.

I therefore conclude that I cannot concur with MacDonald J on this point.

I should deal with one further point not raised in argument, but which has caused me to reflect further on the result. That point is that 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. Here too there are very good reasons why this was not the first or second iteration of the argument. As I have noted above, the argument before the Divisional Court was necessarily predicated on a pre-Supreme Court in Arch approach to causation. The current iteration was pointless at that stage. In Brushfield there was focus on other areas of the policy – and in particular an argument as to the ambit of the MSDE clause which logically prevented taking this particular iteration. I am satisfied therefore that the route the point has taken to emerge does not itself stand in the way of the conclusion I have reached being correct.

In my judgment an orthodox approach to construction points to the conclusion that this clause provides a localised cover, but one which is capable of extending to disease. The natural reading of the clause is that the driver of the clause is the result, not upon the nature of the danger which produces it. On that basis it is consistent with the conclusion which the Supreme Court reached on disease clauses that there is no reason why that danger cannot be one or more cases of COVID-19 within the radius. For the reasons I have given I see no reason why that approach is not applicable, once it is established that disease can be a danger.

This therefore deals with construction.

However the second stage of the argument is causation. Here the question is whether I am bound by the Supreme Court's approach – or if not, whether I am bound (in the other direction) by the approach of the Divisional Court.

I consider that while I may not be technically bound by the Supreme Court's judgment on this point, the logic of the judgment, particularly in its treatment of the hybrid and Arch prevention of access clauses, effectively undercuts the Divisional Court's approach and erodes a principled distinction to vanishing point. While the Supreme Court placed weight on certain aspects of the factual background to the disease clauses, such as the rapid and unpredictable spread of notifiable diseases and the likelihood of central government action in response to the same, this provides insufficient basis for drawing a distinction.

The starting point is that the Supreme Court's judgment certainly applies the broader approach to causation to the hybrid/NDDA clauses with which it was concerned. Further it does so with indications that it considers that the analysis would apply elsewhere.

Thus it concludes:

“the public authority clause in the Hiscox policies indemnifies the policyholder … regardless of whether the loss was concurrently caused by other (uninsured but non-excluded) consequences of the Covid-19 pandemic which was the underlying or originating cause of the insured peril.”

On Arch: “… there were concurrent causes of loss, each sufficient to cause loss without the other. Furthermore, these concurrent causes arose out of the same underlying or originating cause, namely the Covid-19 pandemic. As with the Hiscox policies, on the correct interpretation of the Arch wording, such loss is in our view covered by the policy.”

As to RSA: “ properly interpreted, the clause covers loss caused by the two elements of the insured peril operating in the required causal sequence, but does so regardless of whether any other (uninsured but non-excluded) consequences of the same underlying fortuity (the Covid-19 pandemic) were concurrent causes of the loss.”

The judgment then deliberately considers other similar clauses thus:

There is no reservation, there is no rider. While Axa puts emphasis on the disease element of the hybrid clauses which (SC [213]) “contain, as one element, an occurrence of an infectious disease within a specified distance of the insured premises” this passage explicitly addresses itself not only to hybrid clauses, but also to prevention of access clauses. They are equated and treated on an equal footing. It is unlikely that Lords Hamblen and Leggatt would have missed the point (still less so with the knowledge of the other denial of access clauses very much alive and in conjunction with this use of words).

Axa contends that in relation to the Hybrid and Arch clauses the Supreme Court was “not considering whether, in the case of a non-disease clause providing a narrow, localised cover against incidents specific to the locality of the insured’s premises, it was necessary to show that the event specific to the locality of the insured’s premises was a ‘but for’ cause of action that restricted or prevented access”. That is true. However it is hard to avoid the conclusion that, when referring to other “prevention of access clauses in relation to which, as noted earlier, this issue does not affect the outcome of the proceedings”, the Supreme Court was intending to refer to (inter alia) exactly such clauses. It is notable that the Claimants’ argument makes full contextual sense, discerning here a cross reference to paragraph 122 of the judgment, where the Supreme Court says this:

Axa’s approach, by contrast, cannot link up the cross reference and the issue being referred to in a similarly coherent way.

Further the suggestion that there is a relevant distinction between the hybrid clause and this clause because the hybrid clauses explicitly mention disease is a false point. If on its true construction an NDDA clause does cover disease, even if not explicitly there is no obvious reason why there should be a distinction.

It is certainly true that the Supreme Court reached their conclusion by construing the disease clauses against their factual matrix and that at SC [191-2] stressed the importance of this. The key point is that:

Axa is quite right that in the abstract this leaves the door open to concluding that certain clauses sound in “but for” causation. Axa is also right that at [194] specific factual matrix relevant to disease is cited.

However that begs the conclusion to which the Supreme Court does apparently come. And further this approach seems to circle back to the construction argument – this was apparent in the arguments deployed at paragraph 72 of Axa's skeleton – for example at 72.3 “The NDDA concerns a different insured peril and is clearly focused on circumstances where action is taken in response to a local danger or disturbance specific to the insured premises, not including disease or, alternatively, a range of local events specific to the insured premises of which the occurrence of a disease is only one” and in reliance on the judgment in Brushfield dealing with construction.

Having considered the issue of construction if (and ex hypothesi that is the case) the conclusion is that the clause can cover denial of access arising from disease if the other requirements of the clause are also met, it is hard to see why this reasoning, applicable to disease liability, should not equally apply to coverage for disease accessed via this different portal. For example the underlying academic arguments which form the basis of the Supreme Court’s approach seem to “read across” to this situation just as well as they do to the disease clause.

I therefore conclude that the better argument is that in this context the Supreme Court’s approach to causation should be adopted. I conclude that COVID-19 is capable of being a danger within one mile of the insured premises, which, coupled with other uninsured but not excluded dangers outside, led to the regulations which caused the closure of the businesses and caused the business interruption loss.

In relation to quantum, the Defendant accepts that if the Claimants succeed on the Coverage Issue, they are (subject to proof of their alleged losses) entitled to an indemnity of up to £250,000 in respect of each of the three Restrictions, i.e. a total of £750,000.

The issue dividing the parties is whether, on a proper interpretation of the Policy, the Claimants are entitled to an indemnity of up to £250,000 in respect of each of the three instances where access was restricted for each of the Claimants’ premises, i.e. up to £750,000 for each of their premises, rather than £750,000 in total.

The Claimants submit that:

Cover under the Policy was in respect of “interruption and interference with the business where access to your premises is restricted or hindered.” Each restaurant was a separate business carried on in different premises and each restriction or hindrance interfered with a different business operated by a different insured. Accordingly, each interference led to a different “claim”.

If Axa wished to prevent aggregation in circumstances where the same danger affected different premises in diverse geographical radii, it was open to it to draft the Policy in such a way as to do so. It failed to do so.

Each claim depended on demonstrating action by the police or other statutory body within a one-mile radius. In circumstances where each restaurant was in a different location, each one-mile radius was different and accordingly there was a different claim in respect of each premises.

The Policy contained separate limits for each restaurant in respect of different risks insured. The Policy used the term “floater” in respect of business interruption caused by physical loss and damage, suggesting that there was a floating limit for claims in respect of all the premises for business interruption caused by physical loss and damage. It did not use such a term in respect of NDDA cover so there is no scope for aggregating claims.

The Defendant’s acceptance that there are different claims when the interruption or interference occurs at different times, but not in different places, is illogical and leads to absurdity.

The Claimants also submit that the fact that each claim in respect of each insured, and also each separate premises was “one claim” is emphasised when the composite nature of the policy is considered. The Claimants rely in this regard on the statements by the editors of MacGillivray on Insurance Law and the decision of the Court of Appeal in New Hampshire Insurance Co Ltd v MGN Ltd.

The Claimants accordingly seek (i) a declaration that the Defendant is bound to indemnify them up to a maximum amount of £250,000 per premise for each restriction/closure and (ii) a declaration that the Defendant is bound to pay damages in respect of its wrongful failure to indemnify each Claimant in respect of each premise in respect of each restriction/closure.

Axa submits that:

In contrast to the property damage section, the business interruption section of the Policy does not stipulate separate sums insured specific to particular premises. The use of an aggregate limit suggests that it is intended to apply to all premises and all companies insured; no individual premises or company has the benefit of a specific sum.

The Claimants’ reliance on the fact that the policy is described as “composite” is unavailing because the use of the term does not necessarily mean that there are multiple policies or that the policy is to be treated as if separate policies were issued to each person listed as an insured. Here, “floating” indicates that the sum insured is to be treated as applying across all insured.

The Defendant accordingly denies that the Claimants are entitled to any relief. The Defendant counterclaims for a declaration that it is not liable to indemnify the Claimants, or any of them, in the amounts claimed by the Claimants or any amounts at all. Further or in the alternative, the Defendant seeks a declaration that its liability is limited to the sum of £250,000 for each set of restrictions across all the premises, and £750,000 in total.

There were effectively two strands to the argument on this point. The first was essentially a construction argument, based on the wording of the policy. The second was a legal argument derived from the fact that the policy was a composite policy.

Logically the legal argument, which may be said to inform the basis on which one approaches the document, probably comes first.

As to this, there is no invariable rule. However it is probably fair to say that the expectation raised by the authorities is that a composite policy is treated as a series of contracts – and will hence be treated as giving the relevant cover per contract.

A policy which insures the interests of a number of different insured persons under one policy of insurance is a “composite policy” and takes effect legally by way of separate contracts of insurance between Axa and each of the individual insured companies

The main authorities to which I was referred are:

MacGillivray on Insurance Law 14th Edition, which states at paragraph 1-202:

Arnould on Marine Insurance 20th Edition states at paragraph 8-13: “…a composite policy is treated as a series of bilateral contracts between the underwriters and each assured…”.

MacGillivray paragraph 17-035: “Recent decisions suggest that the courts will readily construe composite policies as containing separate contracts between the insurer and each co-insured”, by reference to Potter J in New Hampshire Insurance Co Ltd v MGN Ltd, [1997] 1 LRLR 24, 41 and per Staughton LJ in the CA at [1997] 1 LRLR 24, 57.

Although Mr Christie made a strong attempt to shift the effect of these authorities, in particular by questioning whether Staughton LJ’s dictum really did support the proposition for which it is cited (also in Clarke on Insurance Contracts 30-4) the central point on which these citations are based remains, and is applicable here.

The central point derives from the judgment of Greene MR in General Accident Fire and Life Assurance v Midland Bank Ltd [1940] 2 KB 388:

It is probably fair to say, as Mr Christie does, that the New Hampshire case in the Court of Appeal does not really take matters much further in that it draws that same distinction between separate interests to insure as opposed to a joint interest before passing on to consider the question of attribution which was key in that case. More to the point is perhaps the judgment of Potter J at p. 41 of the report, where he said:

It is also fair to say that the conclusion may well turn on the facts of the particular case. Axa referred me to the judgment of the Supreme Court of New Zealand in New Zealand Fire Service Commission v Insurance Brokers Association of New Zealand Inc [2015] NZSC 59, at [107-154]. That is a lengthy judgment (running to 158 paragraphs) concerning a policy of insurance which was issued to the New Zealand port collective, which comprised eight separate ports within New Zealand, the point being whether a government body was entitled to levy eight charges or only one. The Supreme Court reached the conclusion that there were eight separate contracts in circumstances where the different ports were separately invoiced for the premium, there was a specific provision providing for construction of the policy as if issued separately to each insured and where the cancellation clause also tended to suggest separate policies. However, as Mr Christie fairly said in the course of submissions the final factor relied on by the court appears to have been the driver – the policy in favour of the universality of the levy (or maximising revenue). In those very particular circumstances I do not gain much assistance from this case. I do not consider that the case establishes these indicia as a test and I do not consider the absence of those indicia in this case is significant.

The exercise to be performed is to ascertain whether the interest is joint or not. Asking that simple question, the answer is perfectly straightforward: it is not. Each company has a separate interest represented by the restaurant or restaurants/café which it owns (Footnote: ⁵). The policy therefore falls to be analysed as a composite policy.

That is not an insignificant conclusion because although it is not beyond the bounds of possibility that there could be a composite policy with a single limit which applies to all the premises and all the claims, that would certainly not be the expectation in the context of a composite policy.

One then moves on to the construction points. Here the most powerful points – and the ones which unequivocally supports the Claimants' position, are the facts that:

The Policy refers to cover in respect of “interruption and interference with the business where access to your Premises is restricted …”;

The premises were in different locations and could well be differently affected by a danger triggering cover. Mr Gruder's nuclear incident in Central London would leave Café Wolseley at least untouched. Further, as he pointed out, closures from two suspicious vehicles (one near the Delaunay and one near the Wolseley) must be seen on any analysis as two separate incidents which would naturally give rise to two claims; and there is no logical distinction if it is the same car, equidistant from the two venues which closes both premises. 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. That then harmonises with the fact of different named insureds and the separate interests which underpin a composite policy.

The other points I consider to be less telling. Mr Gruder is of course right that Axa could have done better to at least attempt to draft its way out of that conclusion. But equally, as Mr Christie contended, construction by reference to what is missing is not regarded as being a particularly safe guide.

As for the fact that the business interruption section of the Schedule does not stipulate separate sums insured specific to any particular premises of any particular group company (in contrast to the property damage section where separate sums are identified for each premises), this is an argument which cuts both ways. Mr Christie says that the distinction is pertinent, but Mr Gruder contends that since the policies are covering both risks the property damage limits effectively support his position. I would tend to agree with that submission, but would not be inclined to give it very much weight.

As for Axa's submission that the fact that the NDDA cover limit is expressed to be “100% of the sum insured or £250,000 whichever is less” supports a single limit because it would be inconsistent for the total sum insured to be a floating sum applicable to all the premises and all the companies in aggregate, but the figure of £250,000 to be applicable to each premises, I would regard it as a considerable stretch to see this as being significant. This reads rather as a point where figures have been filled into boilerplate without due thought. And there may well in such a policy be areas where there is per claim coverage and others where there is not.

Overall the picture which emerges from a consideration of the wording and a consideration of the nature of the Policy persuades me without difficulty that the correct answer is that this is a composite policy in respect of which each insured is entitled to claim £250,000 in respect of each claim.

It follows that the claim of the Second, Third, Fifth, Seventh, Ninth and Tenth Claimants succeeds and they are entitled to the declarations which they seek, namely:

That Axa is bound to indemnify each of those Claimants in respect of each of their premises up to a maximum amount of £250,000 in respect of each of the March 2020 closure, the September 2020 restriction, and the November 2020 closure;

Alternatively, that Axa is bound to pay damages in respect of its wrongful failure to indemnify each such Claimant in respect of each of those Claimants’ premises up to a maximum amount of £250,000 in respect of each of the March 2020 closure, the September 2020 restriction, and the November 2020 closure.