[Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback]
	England and Wales Court of Appeal (Civil Division) Decisions
You are here: BAILII >> Databases >> England and Wales Court of Appeal (Civil Division) Decisions >> Liberty Mutual Insurance Europe SE & Ors v Bath Racecourse Company Ltd & Ors [2025] EWCA Civ 153 (21 February 2025) URL: http://www.bailii.org/ew/cases/EWCA/Civ/2025/153.html Cite as: [2025] EWCA Civ 153

[New search] [Printable PDF version] [Help]

IN THE COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM THE HIGH COURT OF JUSTICE
BUSINESS AND PROPERTY COURTS OF ENGLAND AND WALES
COMMERCIAL COURT (KBD)
Mr Justice Jacobs
[2024] EWHC 124 (Comm)

		Rolls Building, Royal Courts of Justice Fetter Lane, London, EC4A 1NL
		21/02/2025

This judgment was handed down remotely at 10.30am on Friday 21 February 2025 by circulation to the parties or their representatives by e-mail and by release to the National Archives

Sir Julian Flaux C:

Introduction

1. These five appeals are against the decision of Jacobs J dated 26 January 2024 in respect of several business interruption insurance claims arising out of the Covid-19 pandemic which were case managed and tried together by the judge in the Commercial Court. The judgment addressed a number of preliminary issues. The insurers' appeals relate to various issues concerned with policy limits in two of the relevant policies. The insureds' appeals (hereafter described as "the furlough appeals") concern whether payments under the Government Coronavirus Job Retention Scheme ("CJRS"), so-called "furlough payments", fall to be deducted from the indemnity payable to the insured by reason of the "savings clause" in each policy.
2. The appeals which were before the Court originally included appeals by other insureds to whom the judge's judgment related, Liberty Retail Limited, Hollywood Bowl Group plc and Fuller, Smith & Turner plc, as well as appeals by insurers against those insureds and Gatwick Investment Limited (who remain appellants in the furlough appeals before the Court) but those appeals have all been withdrawn. In addition, insurers withdrew appeals on the issue of causation in the Starboard and Gatwick cases. It will not be necessary to say much further about these matters.

Factual and procedural background

3. The Starboard claimants are twenty-one companies, each of which owns or operates a separate hotel in England. The Gatwick claimants are six insured companies each of which owned or operated a separate hotel in England. The Bath Racecourse claimants are twenty companies in the Arena Racing Group which owned or operated racecourses and related facilities such as greyhound tracks, golf courses, hotels and a pub, at twenty-one locations in England and Wales and two group companies operating across the group without their own locations. The main insurer defendant is Liberty Mutual. It, together with Allianz and Aviva, insured the Bath Racecourse claimants under a policy on the so-called Bluefin/Liberty wording for the period from 1 January 2020 to 31 December 2020. Liberty Mutual alone insured the Starboard claimants under a Commercial Combined Policy from 1 July 2019 to 30 June 2020. Liberty Mutual alone insured the Gatwick claimants under a Commercial Combined Policy on essentially the same wording as the Starboard policy from 8 October 2019 to 20 October 2020.
4. The history of the outbreak of Covid-19 and the Government response is set out by the judge in detail in [16] to [48] of his judgment. None of the parties took issue with that analysis. For the purposes of the present appeal, it is only necessary to record the following matters. From 21 March 2020, restaurants, cafes and bars were closed pursuant to the 21 March 2020 Regulations. Those Regulations were revoked by the 26 March 2020 Regulations which imposed more stringent restrictions, the effect of which, so far as the Starboard and Gatwick claimants are concerned, was that their hotels were prohibited from receiving guests other than in very limited exempted categories. Restaurants and bars were closed and any residents who could lawfully stay in a hotel had to be served meals in their rooms.
5. The 26 March Regulations were revoked and replaced with more limited restrictions on 4 July 2020. Although hotels and the restaurants and cafes within them could legally reopen, there were strict social distancing and cleansing requirements. From 14 September 2020, the September Regulations introduced the Rule of 6 prohibiting meeting socially in groups of more than six, including in hospitality venues. Further Regulations on 18 September applied the Rule of 6 more strictly requiring an "appropriate distance" between tables and from 24 September a 10pm curfew was imposed at hospitality venues other than in respect of online deliveries.
6. So far as the CJRS is concerned, this was first announced by the Chancellor of the Exchequer, Rishi Sunak, in a speech on 20 March 2020, when he said:

"This week, the Government has taken unprecedented steps to fight the coronavirus. We have closed schools. We have told people to stay at home to prevent the spread of infection. We are now closing restaurants and bars. Those steps are necessary to save lives. But we don't do this lightly - we know those measures will have a significant economic impact. I have a responsibility to make sure we protect, as far as possible, people's jobs and incomes. Today I can announce that, for the first time in our history, the government is going to step in and help to pay people's wages. We're setting up a new Coronavirus Job Retention Scheme."

7. As the judge noted at [399] and [401] of his judgment, on 23 March 2020 the Coronavirus Bill was debated in Parliament and on the same day the Government published a news story on the CJRS. On 26 March 2020, the Chancellor of the Exchequer gave a further speech in which he discussed the CJRS and the Government published guidance on the CJRS and how to make an application.
8. The CJRS was enacted on 15 April 2020, pursuant to sections 71 and 76 of the Coronavirus Act 2020, by a Treasury Direction which was updated by similar Directions thereafter. The effect of the Scheme was described by the judge at [404] and [405] of his judgment in these terms:

"In overview, under the CJRS, until 30 September 2021 (when the scheme ended) UK employers could make a claim to obtain payment / reimbursement from HMRC of up to 80% of expenditure incurred on costs of employment of qualifying "employees" who were not working but kept on payroll (i.e. "furloughed") for more than 21 days (before 30 June 2020) by reason of circumstances arising as a result of coronavirus or coronavirus disease ("furloughed employees"), up to a maximum of £2,500 a calendar month per employee. Reimbursement of employer expenditure (including expenditure on employer national insurance contributions and pension contributions) was to be made by HMRC if the conditions of the scheme were satisfied.

405.         The basic approach of the CJRS was, therefore, to reimburse employers for the continued payment of furloughed workers. Thus, employees were "furloughed" for the purposes of the CJRS if they were put on a period of leave during which they were instructed to cease all work for the employer in accordance with the CJRS, and employers recovered reimbursement of pay from HMRC in respect of furloughed employees."

9. He then set out the terms of the Treasury Direction at [406]. For the purposes of these appeals, it is only necessary to highlight a few of the provisions of the Treasury Direction:

"2.1 The purpose of CJRS is to provide for payments to be made to employers on a claim made in respect of them incurring costs of employment in respect of furloughed employees arising from the health, social and economic emergency in the United Kingdom resulting from coronavirus and coronavirus disease.

2.2 Integral to the purpose of CJRS is that the amounts paid to an employer pursuant to a claim under CJRS are only made by way of reimbursement of the expenditure described in paragraph 8.1 incurred or to be incurred by the employer in respect of the employee to which the claim relates.

6.1 An employee is a furloughed employee if: a) the employee has been instructed by the employer to cease all work in relation to their employment, b) the period for which the employee has ceased (or will have ceased) all work for the employer is 21 calendar days or more, and the instruction is given by reason of circumstances arising as a result of coronavirus or coronavirus disease.

8.1 Subject as follows, on a claim by an employer for a payment under CJRS, the payment may reimburse: a) the gross amount of earnings paid or reasonably expected to be paid by the employer to an employee b) any employer national insurance contributions liable to be paid by the employer arising from the payment of the gross amount c) the amount allowable as a CJRS claimable pension contribution.

8.2 The amount to be paid to reimburse the gross amount of earnings must (subject to paragraph 8.6) not exceed the lower of: a) £2,500 per month, and b) the amount equal to 80% of the employee's reference salary (see paragraphs 7.1 to 7.15)." (my underlining)

The relevant terms of the insurance policies

10. The Bath Racecourse Bluefin wording under Risk Details Section 2 Business Interruption gave three sums insured for Estimated Gross Revenue with three different Maximum Indemnity Periods: £66,656,147 with a Maximum Indemnity Period of 12 months, £16,466,592 with a Maximum Indemnity Period of 24 months and £25,515,911 with a Maximum Indemnity Period of 36 months. These were different total or aggregate limits for respectively racetracks, golf courses and hotels.
11. The relevant cover for present purposes was under Section 2, headed "Particular Settlement Terms", in respect of Denial of Access ("DOA"):

"Denial of Access

This Section extends to include any claim resulting from interruption of or interference with The Business carried on by The Insured at The Premises in consequence of

…

(b) action by the Police Authority and/or the Government or any local Government body or any other competent authority following danger or disturbance within a one-mile radius of The Premises which shall prevent or hinder use of The Premises or access thereto

…

provided that after the application of all other terms conditions and provisions of this Section the liability of the Insurer shall not exceed

…

(i) GBP 1,000,000 in respect of (a) above any one loss

(ii) GBP 1,000,000 in respect of (b) above any one loss"

12. Under "Conditions" in the Schedule it was provided that the Bluefin wording was "amended as follows". Condition 22 then provided:

"22. Notwithstanding anything contained herein to the contrary, the limit in respect of Section 2 - Particular Settlement Terms, Denial of Access:-

Proviso (i) is amended in respect of (a) to GBP 1,000,000 and a maximum indemnity period of 3 months.

Proviso (ii) and (iii) are amended in respect of (b) and (c) to GBP 2,500,000 and a maximum indemnity period of 3 months."

13. Section 2 of the Bluefin wording also included a Savings Clause in these terms:

"Savings

If any of the charges or expenses of The Business payable cease or reduce in consequence of the Damage such savings during the Indemnity Period shall be deducted from the amount payable."

14. The Bluefin wording also contained a Claims Preparation Clause ("CPC") in these terms:

"Claims Preparation Clause 

Notwithstanding anything contained herein to the contrary this Certificate is extended to pay the exceptional costs not otherwise covered herein necessarily and reasonably incurred by the Insured with the Insurer's prior consent to prepare and verify the amount of claims admitted under this Certificate in accordance with the claims conditions of this Certificate where such claims are in excess of GBP 50,000 above the applicable deductible.

These costs shall not include the cost of negotiation of the claim with the Insurer or its representatives.

The liability of the Insurer under the terms of this Condition shall not exceed GBP 50,000 in respect of any one claim or series of claims arising from a single occurrence."

15. The Starboard Policy provided in the Schedule that the Insured was "Starboard Hotels Ltd & Associated Companies". Those companies were then all set out under "Named Insureds" and "Additional Named Insureds" by endorsements effective from inception and the 22 individual hotels/hotel sites and two offices were identified as The Premises in a schedule. The relevant cover was under a "Prevention of Access (Non-Damage)" or "POAND" extension in these terms:

"Under Business Interruption loss following interference with the Business carried out by the Insured in consequence of action by the Police or other Statutory Authority following danger or disturbance within 1 mile of the Premises which shall prevent or hinder use of the Premises or access thereto or, interference with the Business carried out by the Insured.

Provided that the Company shall not be liable under this extension for more than the amount shown against this extension in the Schedule.

Subject to the terms, Conditions, limits and Exceptions of this Policy"

16. The Schedule to the Starboard Policy set out at Section 2 the Basis of Cover for Business Interruption giving Declared Values for Gross Revenue including Increased Cost of Working of £67,451,597 and under the column headed "Limit of Indemnity GBP" a figure of £89,933,214 with, in the next column, a Maximum Indemnity Period of 24 months. With other covered items, a Total Business Interruption figure of £106,634,638 was given under the "Limit of Indemnity GBP" column. There were then Business Interruption Extensions set out with columns headed "Limit GBP" with another column headed "Maximum Indemnity Period (months)." The relevant extension was number 15, "Prevention of Access (Non-Damage)" which under the "Limit GBP" column provided for a limit of £1,000,000 and under the "Maximum Indemnity Period" column a figure of 3 months.
17. Section 2 of the Policy set out the terms of Business Interruption insurance. By an amendment from inception, the basis of cover was "Gross Revenue including Increase in Cost of Working - Declaration Linked Basis", the underlined words being the savings clause in this Policy:

"Under Business Interruption the insurance under this item is limited to a) Loss of Gross Revenue and b) Increase in Cost of Working and the amount payable as indemnity thereunder shall be:

a) In respect of Loss of Gross Revenue, the amount by which the Gross Revenue during the Indemnity Period shall fall short of the Standard Gross Revenue in consequence of the Incident;

b) In respect of Increase In Cost of Working the additional expenditure necessarily and reasonably incurred for the sole purpose of avoiding or diminishing the reduction in Gross Revenue which but for that expenditure would have taken place during the Indemnity Period in consequence of the Incident but not exceeding the amount of reduction in Gross Revenue thereby avoided;

less any sum saved during the Indemnity Period in respect of such of the charges of the Business payable out of Gross Revenue as may cease or be reduced in consequence of the incident.

Notwithstanding anything herein contained to the contrary the liability of the Company shall in no case exceed: in respect of Gross Revenue 133.33% of the Estimated Gross Revenue stated herein; in respect of each other Item 100% of the Sum Insured; or in the whole the sum of 133.33% of the Estimated Gross Revenue and 100% of the Sums Insured by other Items."

18. "Limit of Indemnity" is defined in the Definitions as follows:

"Limit of Indemnity shall mean:

(a) for the purposes of Sections 1 to 6, the total liability of the Company for all amounts payable in accordance with the Insuring Clause under these Sections for any loss or series of losses arising from any one occurrence as stated in the Schedule. For the avoidance of doubt the Limit of Indemnity is inclusive of the relevant Deductible stated in the Schedule…"

19. Since the only issue remaining in the appeals between the Gatwick claimants and the insurers is the one concerned with furlough payments it is only necessary to record that the Basis of Cover and savings clause in the Gatwick Policy wording were in identical terms to the provisions in the Starboard Policy wording set out at [16] above. The figures for Limit of Indemnity and Limit in the Schedule to the specimen policy before the Court were lower than in the case of Starboard not least because (since the Gatwick claimants were insured under separate policies) it related only to the Crowne Plaza at Gatwick Airport. The maximum indemnity period for the POAND cover was also 6 months not 3.

The judgment below

20. In relation to the insurers' appeals, given that these are in a narrow compass, it is only necessary to focus on the judge's reasoning in relation to the outstanding issues, which concern (i) whether the £1,000,000 limit in the Starboard Policy is applicable separately to each insured under what it is common ground is a composite insurance or, as insurers contend, an aggregate limit; (ii) whether the DOA limit in the Bath Racecourse Policy as amended by Condition 22 is £2,500,000 "any one loss" or an aggregate limit for all DOA claims of £2,500,000. The insurers also argued, as in Starboard, that although the Bath Racecourse Policy was a composite policy, a single aggregate limit applied to the claims by each insured under the individual contracts of insurance contained within the composite policy; and (iii) whether the £50,000 in the Claims Preparation Clause in the Bath Racecourse Policy is any one loss or an aggregate limit.
21. Some of the matters relevant to the insurers' appeals were addressed by the judge in the section of his judgment dealing with Gatwick. At [219] and following the judge dealt with the issue whether "Limit" in the Schedule under the "Business Interruption Extensions" meant "Limit of Indemnity". He preferred the submissions of Mr David Scorey KC for the insurers that there was no distinction between "Limit" and "Limit of Indemnity". For present purposes, it is only necessary to cite [230] of the judgment, where the judge concluded:

"…I do not consider that the reasonable reader would conclude that there was any fundamental distinction between "Limit" and "Limit of Indemnity", whereby the former but not the latter resulted in per interference/ per premises cover. Rather, "Limit" is indeed simply a shorthand for "Limit of Indemnity". Thus, in the context of the £ 1,000,000 POAND cover (and indeed the other covers listed under "Limit" in the policy Schedule Section 2) the aggregation provisions in the definition of "Limit of Indemnity" apply equally to "Limit"."

22. The judge then went on to consider at [232] and following whether the POAND endorsement provides for an annual aggregate limit for the POAND cover. The insurers' argument was based upon the final words of the endorsement:

"Provided that the Company shall not be liable under this extension for more than the amount shown against this extension in the Schedule".

23. The judge rejected the insurers' argument, saying at [233]:

"233. In my view, this argument is effectively destroyed by the success of the argument of insurers which is considered above: i.e. that the "Limit" in the policy Schedule Section 2 was to be equated with "Limit of Indemnity" as defined in the policy. The "Limit" is therefore, by the express terms of the policy, an "any one occurrence" limit. I do not consider that the words of the POAND endorsement can reasonably be read as imposing any limit beyond the "any one occurrence" limit which is thus provided for in the policy Schedule. The final words of the endorsement simply mean that Liberty Mutual and Aviva are not liable under the POAND extension for more than £ 1,000,000 for any loss or series of losses arising from any one occurrence."

24. The judge went on to also dismiss the insurers' argument that there was an annual aggregate limit for the POAND cover by virtue of the proviso in the Insuring Clause that their liability shall not exceed "the aggregate Limit of Indemnity as stated in the Schedule" holding at [240] that the provisions in the Insuring Clause could not readily be applied to the POAND extension. At [246] he concluded:

"Accordingly, I reject the argument that the POAND clause is subject to an aggregate limit. In my view, the relevant limit is "any one occurrence" as provided for in the Limit of Indemnity provision, and there is no aggregate limit."

25. At [267] the judge set out the submission by Mr Gruder KC that the policy limits in the Starboard Policy applied separately to each insured company, relying on the decisions of Potter J and the Court of Appeal in New Hampshire Insurance Co Ltd v MGN Ltd [1996] CLC 1692; [1997] IRLR 24 ("New Hampshire") and of Cockerill J in Corbin & King Ltd v Axa Insurance UK plc [2022] EWHC 409 (Comm); [2023] 1 All ER (Comm) 429 ("Corbin & King"). He submitted that there was no warrant for reading the limit as applying to all of the premises, contrary to the expectation of a composite policy. The POAND Limit applied separately to each of the Starboard insureds and consequently each of the premises under the separate contracts of insurance comprised in the composite policy. Were it otherwise, the sensible commercial decision of related companies insuring their respective interests under one policy document would be a trap for the unwary. It made no sense to say Starboard was in a worse position than Gatwick because they had adopted the convenient route of having a single document rather than separate documents for each insured.
26. At [268] the judge recorded the submission of Mr Scorey KC for the insurers that there were a number of serious flaws in the reasoning of Cockerill J in Corbin & King. She had started in the wrong place by attaching significance to a legal argument based on the nature of a composite policy, whereas the only relevant question was how the contract was to be construed, applying ordinary principles of construction. Contrary to her conclusion, there was no "expectation" that each policy would have separate limits by virtue of the policy being composite. She was wrong to place reliance on the decision in New Hampshire, which took the analysis no further. He submitted that insurers could not sensibly be expected to rate risks by analysing whether the insurable interest of the co-assureds are, as a matter of law, to be treated as several and distinct, which is the test for whether the policy is joint or composite. Furthermore, if ordinary commercial policyholders were to focus on the limit set out in the policy, they would naturally read that limit as just that: the limit and on an aggregate basis.
27. The judge accepted Mr Gruder KC's arguments saying at [273]:

"In Corbin & King, Cockerill J concluded "without difficulty" that the correct answer was that the policy in that case was a composite policy in respect of which each insured could claim up to the relevant policy limit. I do not consider that there is any material distinction, in that context, between the composite policy at issue in Corbin & King and the composite policy covering the various Starboard insureds."

28. At [274] he summarised Cockerill J's reasons for reaching that conclusion:

"Her reasons for reaching her conclusion were in summary as follows. The policy was a composite policy, covering insureds with separate interests to insure. It was not therefore a policy covering joint interests in the same property. Whilst there was no invariable rule, it was fair to say that the "expectation raised by the authorities is that a composite policy is treated as a series of contracts - and hence will be treated as giving the relevant cover per contract". Each company had a separate interest represented by the restaurant or restaurant(s)/café(s) which it owned, and the policy therefore fell to be analysed as a composite policy. That was:

"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"."

29. He noted at [275] that the wording in Corbin & King provided cover for interruption and interference with the business "where access to your Premises is restricted". The premises were at different locations and could well be differently affected by a danger triggering cover. He said at [276] that these considerations applied equally to the Starboard Policy and he was unpersuaded by the arguments that Cockerill J's reasoning was flawed. He regarded it as amply supported by the decision in New Hampshire as well as the major textbooks.
30. At [278] he concluded:

"Furthermore, I do not consider that there is any material distinction in the wordings which would lead the court to reach a different conclusion to that reached by Cockerill J. The POAND endorsement in the present case refers to "Business Interruption loss following interference with the Business carried out by the Insured in consequence of action …. following danger or disturbance within 1 mile of the Premises which shall prevent or hinder use of the Premises". Just as in Corbin & King, each of the Premises owned by each of the Starboard claimants was in a different location and could well be differently affected by a danger triggering cover. In the context of a composite policy covering the separate interests of each named insured, the limit in the POAND endorsement is sensibly to be construed as applying separately to each named insured. I accept, as did Mr Gruder, that it would be possible for a composite policy to provide for what could be called a "shared" limit. However, I see nothing in the language of the policy, or its context, which points in that direction. On the contrary, I consider that a reasonable policyholder, knowing that each hotel was owned by a separately named insured, would conclude that the £1,000,000 limit applied to each insured in respect of an interference which might affect that insured, and would not understand it as creating a shared limit."

31. At [279] he reached the same conclusion about the absence of an aggregate limit in the POAND clause in the Starboard policy as he had reached at [246] in the case of the Gatwick policy (set out at [24] above).
32. The judge set out the parties' arguments on the second issue concerned with the Bath Racecourse Policy at [363] and following. He noted the contention of the Bath Racecourse claimants that the original DOA cover was on the basis of "any one loss" and the effect of the amendment was to raise the limit to £2.5 million and add a bespoke Maximum Indemnity Period. They submitted that the insurers' contrary argument had a number of insuperable problems. The amendment does not provide a full replacement text or state that the words in the proviso are "hereby deleted". The reasonable reader would understand the same basis (any one loss) to be intended. Where the parties intended to move to a limit in the aggregate in the Conditions in the Policy, they expressly said so, as in Condition 7 for material damage to golf greens ("subject to a limit of GBP 20,000 in the aggregate"). To the extent there was genuine ambiguity on the point, it would be appropriate to construe the Policy contra proferentem the insurers. If for some reason "any one loss" has been impliedly struck through, the default position for limits is that they apply per occurrence, because the wording in relation to Excesses is: "All claims for Damage arising out of one occurrence … shall be adjusted as one claim and from such adjusted claim the sum specified below shall be deducted".
33. At [365] the judge recorded that it was common ground that the Policy was a composite contract of insurance, so the Bath Racecourse claimants adopted the submissions of other insureds that the applicable limit of £2.5 million applied, at least, on a per claimant basis. They also submitted that a particular claimant might have more than one loss within the "any one loss" language, for example where it had more than one affected premises but, as the judge said, that point was to be determined at a later stage.
34. At [366] the judge set out the insurers' argument that the "any one loss" language in limb (b) of the DOA Clause had been replaced in its entirety. That must be the effect of Condition 22 which makes no reference to "any one loss". The amendment of proviso (i) makes little sense if the words "any one loss" were intended to be retained, because the limit in the DOA Clause was already £1,000,000 and Condition 22 applies: "Notwithstanding anything contained herein to the contrary".
35. The insurers also argued that, although the Bath Racecourse Policy was composite, it did not in and of itself entitle the Bath Racecourse claimants to one or multiple limits per policyholder. This was essentially for the same reasons as argued in relation to the Starboard Policy. Mr Scorey KC submitted that the Bluefin wording made abundantly clear that: "Unless stated otherwise the Insurer will not pay more than the Sum Insured Compensation or Limits of Indemnity in any one Period of Insurance". He submitted that this meant that the default position is that the limit is the limit for the policy period and not each and every loss or occurrence.  Accordingly, the limit under the DOA Clause is £2,500,000 in total/the aggregate for all of the policyholders under the Bath Racecourse Policy for the period of the insurance.
36. At [368] the judge accepted the Bath Racecourse claimants' argument that there was no change in the "any one loss" language in proviso (ii) in the DOA Clause and that Condition 22 had increased the limit to £2.5 million and added a maximum indemnity period, but not altered the existing agreement as to "any one loss". If it were to disappear, one would expect language such as in Condition 20 which referred to a provision being "deleted and replaced". He considered the more natural reading of Condition 22 is that the relevant amendments were spelt out, apart from which the provision remained as agreed.
37. At [369] he also accepted the argument of Mr Adam Kramer KC for the Bath Racecourse claimants that if the parties had been intending to delete the "any one loss" basis for the original £1 million limit, one would expect them to identify the new basis on which the £2.5 million limit was to operate. He said that Mr Kramer KC had made effective points about Conditions 6, 7 and 8 where there were changes in the basis of aggregation in respect of certain aspects of Material Damage and that there was force in the point based on Condition 7, that if the parties had been intending to introduce an aggregate limit rather than "any one loss", they would have spelt that out.
38. The judge considered these arguments more powerful and persuasive than those of the insurers. He did not consider that the opening words of Condition 22: "Notwithstanding anything contained herein to the contrary" are equivalent to language deleting and replacing all the text of proviso (ii). They have to be read in the light of the fact that that the clause is only making an amendment to that proviso (as well as (i) and (ii)) so the opening words make it clear that the amended proviso applies even if there are other provisions to the contrary.
39. At [371] the judge referred to Mr Scorey KC's submission that the amendment to proviso (i) made little sense if the words "any one loss" were being retained because the limit in the DOA clause was already £1 million. The judge did not think the draftsperson's decision to repeat that the limit was £1 million had the significance ascribed to it by Mr Scorey KC. Arguments based on surplusage are generally weak in the context of commercial contracts like this. If the intention had been to delete "any one loss" in proviso (i) one would expect that to be done expressly and clearly. Where a maximum indemnity period was being added, it was not surprising that the draftsperson decided to make it clear that the overall limit remained where it was. The judge concluded that the relevant policy limit under the DOA clause is £2.5 million for "any one loss". He was not deciding at present how many losses there were.
40. The judge dealt with the argument on the Claims Preparation Clause ("CPC") in the Bath Racecourse policy at [379]. He noted that the only question was whether the £50,000 limit which was in respect of any one claim or series of claims arising from a single occurrence applies in the aggregate across all the claimants or is a limit available to each claimant. The judge noted that this depended upon the composite policy argument which he had resolved in favour of the claimants.
41. The judge then turned to the furlough payment issue which was whether the insureds had to give credit for any payments they received under the CJRS. He set out the factual background to the CJRS which I have summarised at [6] to [9] above. He then referred to the decision of Butcher J in Stonegate Pub Company Ltd v MS Amlin Corporate Member Ltd and others [2022] EWHC 2548 (Comm); [2023] 1 All ER (Comm) 981; [2023] Bus LR 28; [2023] Lloyd's Rep IR 672 ("Stonegate") noting that the identical issue about furlough payments was considered by that judge at [250]-[289]. Butcher J decided that CJRS payments were to be taken into account under a savings clause which provided:

"Any costs normally payable out of Turnover (except depreciation) as may cease or be reduced during the Indemnity Period as a consequence of the Covered Event".

Jacobs J considered that clause was indistinguishable from the clauses in the present case.

42. He noted at [410] that it was common ground in Stonegate that the employment costs were normally payable out of "Turnover". The central issue Butcher J had to consider was whether the CJRS grants had caused the relevant employment costs to "cease or be reduced". There was no dispute that if they had, that was a consequence of a "Covered Event" under that policy. The contention of Stonegate was that there had been no reduction in the employment costs. It had continued to pay wages and had had to do so in order to benefit from the CJRS. Jacobs J noted at [412] that Butcher J had rejected that contention at [258] of Stonegate:

"In my judgment, employment costs were at least 'reduced' pro tanto by reason of the payment of corresponding amounts under the CJRS. I consider that the natural meaning of the definition, including its savings clause, is that it is referring to costs to the business. Insofar as such costs were defrayed by the government, I consider that they were 'reduced'. That, in my view, reflects the net financial effect of payments under the CJRS and the commercial reality."

43. At [412] the judge noted that Butcher J had gone on to identify three further considerations supporting that conclusion which in summary were as follows. First, the applicable accounting standards would permit CJRS payments to be presented as an offset against employment expenses. Second, the scheme envisaged that the Government might make CJRS payments before wages were paid. Butcher J considered that the question whether CJRS payments were to be taken into account under the savings clause could not depend on whether the payments were received before or after payment to the employee. Third, Butcher J considered that the relevant provision should be construed to accord with the basic principle that the policy was a contract of indemnity, citing my judgment at first instance in Synergy Health v CGU Insurance [2010] EWHC 2583 (Comm); [2011] Lloyd's Rep IR 500 ("Synergy"). Butcher J concluded: "CJRS payments did reduce costs payable out of Turnover and are to be taken into account under the savings clause".
44. The judge went on at [413] to refer to Butcher J's consideration of whether the insurers would as a matter of general law have been subrogated to payments of CJRS, part of his judgment which was obiter since he had already decided the savings clause did apply. Butcher J decided the general law could not be relied upon to produce a different result to that specifically provided for. He held, after a full discussion of the relevant authorities, that the general law would produce the same result as the savings clause.
45. At [415] the judge noted that Butcher J had given permission to appeal and that, at the time of the hearing before him, the appeal was scheduled to be heard a few weeks later, but the appeal had subsequently been compromised.
46. At [416] the judge noted that the central argument of Mr Gruder KC was that Butcher J had not had to address the causation question of whether the reduction in costs was a consequence of the insured peril because Stonegate had conceded the question of causation. He submitted that the concession was probably wrongly made and that one issue for the then pending Stonegate appeal was whether the concession could be withdrawn.
47. Mr Gruder KC had submitted that the requirements for CJRS under the Treasury Direction were purely financial. It was irrelevant whether the business had been ordered to close or whether there was Covid-19 at or any particular distance from the premises. The only qualifying condition was that an employer must have a PAYE scheme registered as at 19 March 2020. At [418] the judge noted that Mr Gruder KC submitted that the answer to the question whether the CJRS payments were made as a result of the insured peril, i.e. the action by the Police or other Statutory Authority following danger or disturbance within 1 mile of the Premises which prevented or hindered use of the Premises or access thereto, was obviously no. CJRS was payable to businesses even if the relevant regulations did not cause them to close or interfere with their trade. Payments were not made because the government prevented or hindered access or use of the premises, nor because of a danger (i.e. Covid-19) within 1 mile of the premises in the period leading up to the regulations which imposed restrictions. Mr Gruder KC emphasised that the relevant insured peril in the present case was not simply the disease, but the restrictions imposed in consequence of the disease.
48. The judge noted at [419] that Mr Gruder KC placed reliance, in support of his argument, on the decision of the Full Federal Court of Australia in LCA Marrickville Pty Ltd v Swiss Re International SE [2022] FCAFC 17, at [442]– [463]. There the Court had reversed the trial judge and found that the policyholder did not have to give credit for "JobKeeper" payments. The judge quoted [461] of that judgment and said at [420] that the Court in that case had held that it was necessary for the purposes of the causal requirement in the savings provision "to focus on the criteria for the JobKeeper payments, rather than the general underlying policy of the JobKeeper scheme". Mr Gruder KC submitted that the same approach should be taken here. Focusing on the criteria for payment, the only requirement was a qualifying PAYE scheme. Proof of action by the Police or other Statutory Authority following a relevant danger was not required.
49. The judge then noted that Mr Gruder KC had also referred to the decision of McDonald J in the Irish High Court in Hyper Trust Ltd v FBD Insurance plc [2023] IEHC 455 where the judge had been able to distinguish Marrickville on the basis of differences between the JobKeeper scheme and the relevant Irish schemes, but without any suggestion that Marrickville was wrongly decided. Although McDonald J decided credit had to be given for payments received under the Irish schemes, the criteria for grants under those were not comparable to CJRS.
50. The judge recorded at [423] that Mr Kramer KC supported Mr Gruder KC's argument on causation, explaining that the concession in Stonegate was possibly a consequence of the nature of the insured peril in that case: it was a pure disease cover and applied to the "Vicinity" which would have extended to the whole of the UK. However, the insured peril here was very different, comprising all the elements of the composite peril. Any reduction in wage costs was not a consequence of those elements operating in combination.
51. Unlike Mr Gruder KC, Mr Kramer KC submitted Butcher J's decision in Stonegate was wrong and should not be followed. He had been wrong to decide that there was any reduction in wage costs by reason of CJRS payments. Mr Kramer KC submitted that reimbursement, defrayal and funding of a cost are not reduction of that cost but increases in non-trading income to ensure the business can afford the costs. The judge recorded at [425] his submission that: "Put shortly, paying someone to keep incurring an expense is the opposite of the expense ceasing."
52. At [426] the judge recorded that Mr Kramer KC's submission on causation was somewhat different from Mr Gruder's. He submitted that the "in consequence of" language of the savings clause required proximate causation which meant that it was necessary to find out if the payment was a collateral benefit, as if it was, it would not be in consequence of the insured peril. The important question was whether the CJRS payments were or were to be equated with benevolent gifts. Mr Kramer KC challenged Butcher J's analysis of the general law. He had been wrong to consider the question of collaterality of payments from the perspective of principles of law concerning subrogation, but should have applied a proximate cause analysis and also considered cases outside the insurance context. Butcher J had also been wrong to attach significance to the failure of Stonegate to show that the Government intended to benefit Stonegate to the exclusion of the insurers. The case should be decided on principle with a need to interrogate the character and broader purpose of the payment. Applied here the court should conclude that receipt of CJRS was the same as if the insured had received charitable donations from sympathetic customers, which were collateral payments.
53. At [429] the judge recorded Mr Scorey KC's submission for the insurers that he should follow Butcher J on the "cease and reduce" point. On causation, he submitted first that the savings clause should be approached via the prism of its purpose, which was to avoid over-indemnification. Second, furlough was not simply a gift or donation, but a Government scheme which gave public law rights to employers. The effect of the scheme, if the employer chose to accept the 80% furlough payments was that the employee could no longer work for the employer, in practical terms his time belonged to the Government.
54. Third the scheme was meant to prop up the economy and halt or at least delay redundancies which would have otherwise occurred. The regulations had closed down the economy causing difficulty to businesses which could not afford to pay employees. The scheme was therefore the result of the very peril insured against under the prevention of access clause. It mattered not that furlough payments were available to all employers with a PAYE scheme as all businesses were affected in some way by the restrictions. Fourth, the core element of the peril insured against was the danger or disturbance within the relevant radius. The restrictions were caused by that danger and precisely the same could be said of the furlough scheme. The Government was prompted into action by cases of Covid-19 both inside and outside the relevant radii, all of which had an impact on the economy because of the imposition of restrictions. The scheme was designed to mitigate the effects of the restrictions imposed because of the prevalence of the disease both inside and outside the relevant radius. If a single case within the radius was good enough for the purpose of policyholders establishing concurrent causation, as the Supreme Court had held that it was in Financial Conduct Authority v Arch Insurance (UK) Ltd and ors [2021] UKSC 1; [2021] AC 649 ('the FCA test case'), the same approach should be taken on the other side of the equation with savings.
55. The judge began his own analysis at [435] noting it was common ground that the issue turned on the construction of the savings clauses and Mr Scorey KC did not suggest that if his argument on construction failed the general law of subrogation would produce a different result. The parties accepted the correctness of Butcher J's approach to the general law.
56. The judge considered that the issue of whether the CJRS payments did reduce the relevant costs was precisely the same as considered by Butcher J and he should follow his decision. He was not persuaded by any of Mr Kramer KC's arguments that Butcher J was clearly wrong or wrong at all. He was clearly right. He noted that at [50] of Hyper Trust MacDonald J appears to agree with Butcher J's conclusions on this aspect of the case. The judge considered that Mr Kramer KC's argument substantially repeated points made to Butcher J and rejected by him. One new point was the reliance on the decision of the New South Wales Court of Appeal in Mobis Parts Australia Pty Ltd v XL Insurance Co SE [2018] NSWCA 342; [2019] Lloyd's Law Reports IR 162, which had taken a different approach to depreciation in the context of a savings clause to the approach I took in Synergy.
57. The judge did not regard this as a significant point for a number of reasons. First, Synergy was not critical to Butcher J's analysis. He only referred to it in the context of his third consideration which lent support to his conclusion at [258]. The judge said, at [438]:

"Butcher J relied upon Synergy as further support for the principle that the relevant contractual provision should be construed, if there is any room for argument, to accord with the basic principle that an insurance policy is a contract of indemnity. In that context, Butcher J cited (at paragraph [267]) the judgment of Brett LJ in Castellain v Preston [1883] UKLawRpKQB 69. It is not clear to me that the New South Wales Supreme Court would substantially disagree with Butcher J's proposition."

58. Secondly and in any event, the judge was applying English law, set out in my judgment in Synergy and applied by Butcher J in Stonegate. The judge considered that where there were already two decisions at first instance on a point it should be regarded as settled at first instance and any challenge made on appeal. Accordingly, he rejected Mr Kramer KC's argument that the CJRS payments did not reduce the relevant costs.
59. He then turned to the question of causation, noting at [441] that it was common ground that "incident" in the expression: "in consequence of the Incident" was not confined to "Damage to Property Insured", the definition in the Liberty Mutual wording. It should be read more broadly as a reference to the insured peril.
60. The judge agreed with the insureds' submissions that it is appropriate to look at all aspects of the insured peril. In the context of the prevention of access clauses in issue, this required causation to be considered by reference to all the elements of the composite peril in the clause, not simply to the "danger" element of the peril. However, at [443] he agreed with Mr Scorey KC's submission that it was too narrow an approach to causation to focus only on the question whether, in order to receive a CJRS payment, a policyholder needed to prove those same elements. He agreed that the CJRS cannot be regarded as wholly separated and divorced from the restrictions introduced as a consequence of the widespread prevalence of Covid-19 but on the contrary, they were closely connected. The judge said: "It is obviously no coincidence that the first announcement of the furlough scheme on 20 March 2020 was on the very same day that the government announced that it would be closing down a variety of businesses."
61. The judge said at [444] that the restrictions on and closure of businesses all happened prior to the actual introduction of the CJRS on 15 April 2020 which, he agreed with Mr Scorey KC, was the appropriate date on which to consider causation in the present context. By then the key restrictions relied on by the policyholders here had been introduced. At [445] the judge said:

"It is of course true that the furlough scheme was not simply a consequence of the restrictions on the particular businesses operated by the policyholders in this case. It was a consequence of restrictions which affected a very large number of businesses across the economy as a whole. However, the effect of the decision of the Supreme Court in the FCA test case is that, when considering the operation of the insured peril, a concurrent causation analysis is to be applied. It is therefore sufficient, for the purposes of coverage, for a policyholder to show loss flowing from a combination of an insured peril which affected its business together with similar perils which affected other businesses. I consider that the same approach can and should properly be taken when considering causation in the context of the receipt of CJRS payments. It is therefore sufficient to show that the CJRS (and thus the payments made pursuant to that scheme) was brought into being in consequence of a combination of government restrictions affecting the business of each claimant policyholder in combination with restrictions affecting the business of other policyholders."

62. At [446] he noted that, in the FCA test case, the Supreme Court considered that the overriding principle of considering how the words would be understood by a reasonable policyholder meant that, if possible, the "trends" clauses should be construed consistently with the insuring clauses. The judge considered that a similar approach should be taken in relation to the savings clause. As he then said:

"Thus, as Mr Scorey submitted, the case against the insurers in relation to the peril is a concurrent causation analysis: there was a relevant action by the statutory authority following disease within 1 mile of the premises, and that interfered with the policyholders' business. Equally, the furlough savings were in consequence of what had happened: they were brought in because of damage to businesses caused by the restrictions on a large number of businesses, including those of the claimants, brought in by the government as a result of the pandemic. I agree with Mr Scorey that what works on one side of the line should also work on the other, and that it is not appropriate to take a different and much stricter approach to causation in the context of savings than in the context of the insured peril."

63. The judge did not consider that the decision of the Full Federal Court of Australia in Marrickville dictated any different result. That Court was not considering the factual circumstances of the CJRS and it is by no means clear that a close parallel existed between the factual circumstances in Australia and those in the UK. Also, it does not appear that any argument along the lines of that of Mr Scorey KC, which the judge found persuasive, was advanced. The judge also noted that both Butcher J in Stonegate and MacDonald J in Hyper Trust had distinguished Marrickville. In that case the Court had taken a narrow approach to the causation question by focusing on the criteria for JobKeeper payments, but the judge agreed with Mr Scorey KC that the causation question should not be so narrowly focused.
64. He also noted that MacDonald J had taken a broader approach to the causation question and reached the conclusion that credit for various government payments received by the policyholders should be given. The judge cited various passages from the Hyper Trust judgment, concluding at [451]:

"Accordingly, McDonald J was not focused solely on the criteria for payment under the TWSS, but applied a broader causation analysis. Furthermore, it was no obstacle to the broad causation analysis that the criteria under section 28 of the 2020 enactment applied "more widely than in the context of closures". In other words, an Irish business could obtain TWSS even if its own business had not closed. He also said, in the passage quoted above, that "it could not plausibly be suggested that the closures in place at the time of its enactment (and which were expected to continue thereafter) were not a proximate cause of the TWSS scheme established under the section". It seems to me that the position is the same in the present case in the light of the factual background to which I have referred. Thus, it cannot plausibly be suggested that the closures of the businesses of the various claimants in these proceedings (and which lie at the heart of their claim for indemnity) were not a proximate cause of the CJRS scheme."

65. The judge said at [452] that whilst there were differences between the criteria for payment under the Irish scheme and the CJRS, he did not consider that these affected the causation analysis. At [453] he said that: "I accept Mr Scorey's submission that there is a sufficient and indeed proximate causal connection between the composite insured peril and the CJRS payments which were made and thus reduced the wage costs of the business." He considered that conclusion answered Mr Kramer KC's separate argument on causation which focused on the need for there to be proximate causation between the insured peril and the CJRS payment.
66. He considered that it was therefore unnecessary to deal in any detail with Mr Kramer KC's argument as to the alleged collateral nature of the CJRS payments and his criticism of the judgment of Butcher J. He was unpersuaded that the payments were, or could be equated with, benevolent gifts. They were measures introduced to mitigate the economic effects of the restrictions imposed by the Government. The judge also considered it appropriate to follow Butcher J's decision that the insurers would be subrogated to these recoveries under the general law. Butcher J had considered the leading insurance cases in this area and the judge was not persuaded that his analysis was clearly wrong, or wrong at all. Once the conclusion was reached that an insurer would be subrogated to these recoveries, any argument that they were "collateral" could not be sustained.

The grounds of appeal

67. The grounds of appeal advanced by these insurers for which permission to appeal was granted are set out below. The first ground related to both the Starboard policy and the Bath Racecourse policy, whereas the second and third grounds related only to the Bath Racecourse policy:

(1) The judge erred in law and/or in principle in following the approach of Cockerill J in Corbin & King that the 'expectation' in a composite policy was that it provided multiple limits. Corbin & King was wrongly decided on that point. The judge should have held that, even though composite in nature, each policy in issue contained a single, aggregate limit in respect of all the individual contracts of insurance included with the composite policy, to which a single limit applied.

(2) The judge erred in law and/or in principle in his construction of the Bath Racecourse policy by holding that each claimant is entitled to claim up to the limit of £2.5 million under the DOA clause for any one loss. The judge should have concluded that the amendment to the DOA clause affected both the quantum of the limit and the basis on which it applied, namely from "any one loss" to simply "a maximum indemnity period of 3 months".

(3) The judge erred in law and/or in principle by holding that under the Bath Racecourse policy the CPC cover is limited to £50,000 in respect of any one claim or series of claims arising from a single occurrence. He should have held that the limit is an aggregate limit applicable to the insureds collectively.

68. The grounds of appeal advanced by the insureds on the furlough appeal are that the judge erred in construing and/or applying the savings clauses because:

(1) The insureds' employee costs did not "cease" and were not "reduced" by payments made under the CJRS by the Government;

(2) Payments made under the CJRS were not "in consequence of" the insured peril because such payments did not correlate to the insured peril;

(3) Payments made under the CJRS were not "in consequence of" the insured peril because such payments were collateral/res inter alios acta.

69. The Starboard and Gatwick claimants had an additional ground that there were no savings during the Indemnity Period in respect of the charges of the Business payable out of Gross Revenue.

The parties' submissions on the insurers' appeals

70. It was common ground that the policies in question were to be construed by reference to the understanding of the reasonable policyholder, but Mr Scorey KC for the insurers submitted that one could not imbue the reasonable policyholder with any particular knowledge of insurance law, still less with knowledge of the implications of the policy being composite. He submitted that it followed that the reasonable policyholder would have no a priori expectation as to the number of limits for the basis of cover. However, as Popplewell LJ pointed out in response to this argument, the reasonable policyholder would have a broker to advise as to the effect of the law on composite policies.
71. Mr Scorey KC submitted that the approach in Corbin & King that there was an expectation that there was a separate limit for each insured in the composite policy was unprincipled and unjustified. Although each insured had a separate contract of insurance, the contracts did not operate in isolation but were interrelated, hence having an aggregate limit. The fact that the policies were composite policies did not mean that there were separate limits for each insured. There was no presumption or expectation that there were separate limits simply because it was a composite policy; it all depended upon the construction of the policy. The fact that the composite policy had a limit of £x did not mean that one treated it as a bundle of individual contracts written out with the name of each individual policyholder for the limit of £x.
72. He relied upon the decisions of Jacobs J and of the Court of Appeal in Technip Saudi Arabia Ltd v The Mediterranean & Gulf Insurance and Reinsurance Co. [2023] EWHC 1859 (Comm) and [2024] EWCA Civ 48 ("Technip"). That was a case of a composite policy with an exclusion for property owned by the Principal Assured. KJO, the owner of a wellhead platform, and Technip which had caused damage to the platform were both Principal Assureds under the composite policy. Technip argued that the exclusion did not bite because it was not the owner of the platform. Jacobs J held that the exclusion applied to any property owned by any of the Principal Assureds. That decision was upheld in this Court where Sir Geoffrey Vos MR said at [34]:

"The point about the composite nature of the policy is actually quite a simple one. As already mentioned, reading the policy as a separate insurance for Technip, Technip argued that it was obvious that "the Principal Assured" whose property was referred to in the first limb of endorsement 2 was Technip, not KJO or any other Principal Assured. There is, however, a fatal flaw in that argument. It is true that the policy is a composite policy that is expressly "deemed to be a separate insurance in respect of each Principal Insured". But in reading endorsement 2 in Technip's deemed separate insurance, the words "the Principal Assured" cannot have any different meaning than they have in the other imagined separate insurances for each of the other insureds. Accordingly, if the words "the Principal Assured" mean "Technip and/or KJO and/or associated companies", they must have that same meaning in each separate insurance including Technip's separate insurance."

Mr Scorey KC relied upon that case in support of the proposition that when one is looking at the individual rights of the policyholder in a composite policy who has suffered a particular type of loss covered by the policy, one does not assume that the limit in the policy for that head of loss is specific to that policyholder.

73. Mr Scorey KC referred to the judgment of Cockerill J in Corbin & King at [230] and [238] where the judge said:

"…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…

…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."

Mr Scorey KC submitted that this approach towards a composite policy, that the expectation was that the limits were per insured, was wrong. Rather, it all depended on the terms of the policy.

74. He also referred to the decision of Potter J and the Court of Appeal in New Hampshire relied upon by Cockerill J and by the judge in the present case. That case concerned claims by companies in the Maxwell Group under fidelity policies in respect of liability to pension trustees and others for losses caused by the dishonesty of Robert Maxwell. One of the issues was whether the limits in the policies, which were composite policies, were separate limits for each insured company or was there a single limit applicable to all the insured companies in the group. The relevant limit was £1 million any one loss. As Mr Scorey KC said, Staughton LJ (giving the judgment of the Court of Appeal) dealt with the issue pithily at 63lhc; 1744G saying:

"Despite emphatic protests by Mr Rokison whenever the topic was mentioned, once it had been decided under issue F that each company was separately insured the answer to this question must be that in general there was a separate limit for each company."

75. Mr Scorey KC submitted that this was the bedrock of the idea that composite policies are separate contracts of insurance with their own individual limits "in general". There was no explanation as to why that would be the case in general. He submitted that this conclusion must be read in the light of what Staughton LJ had said slightly earlier in the judgment on this issue:

"This last issue is particularly difficult, largely owing to the obscurity of the contract wording. Despite the assistance of counsel we have found difficulty understanding the questions, let alone deciding them."

76. However, as I pointed out in argument, Potter J does not seem to have had much difficulty deciding the issue. He said at 1726-7; 49 rhc:

"The answer to question (1) seems to me straightforward. Given that I have held that the insurance is composite, then the effect is to create a separate contract of insurance with each insured under which the policy limits apply separately to each company insured."

Mr Scorey KC submitted that this conclusion was not derived from the composite nature of the policy but from the wording of the policy under the terms of which loss to one policyholder would not diminish "any one loss" to another. The conclusion may be right, but he submitted that it was not the foundation for the assertion that there is an expectation of multiple limits simply by dint of the fact that the policy was composite.

77. Mr Scorey KC addressed the submission made by the Bath Racecourse claimants in their skeleton argument that the Court of Appeal in New Hampshire decided that there were individual separate policies within the composite policy each with its own separate limits and that this was part of the ratio and therefore binding on this Court. He submitted that the Court of Appeal had not decided whether the composite policy was one multilateral document or a bundle of individual policies. Dealing with an earlier issue of whether non-disclosure by one insured would constitute non-disclosure by all the insureds, they said at 1737H; 57-8:

"Technically one ought to enquire whether for each layer in each year there was one contract, or as many contracts as there were companies insured. And if the former, can a contract be avoided for non-disclosure as against one or some of the insured, but not against others? We feel that we are relieved from the need to answer those questions by the authority of the House of Lords, in the passage already quoted from P Samuel & Co Ltd v Dumas [1924] AC 431. That, it is true, was not a case of non-disclosure but of wilful misconduct by one of two persons insured. But in our opinion the principle that the innocent party can still recover if it is a separate insurance must equally apply."

78. The reference to the passage from P Samuel & Co Ltd v Dumas [1924] AC 431 was to the passage from Lord Sumner's speech quoted at 1736F to the effect that where there are separate interests, one insured is not affected by the misconduct of another. Mr Scorey KC submitted that the Court of Appeal in New Hampshire had not grappled with the issue whether or not a composite policy is a series of separate contracts of insurance. However, as I pointed out in argument, the Court of Appeal did say at 1737E; 57rhc:

"We agree with the judge that all the contracts of insurance were composite in nature, there being more than one insured 'and each being insured separately."

79. Mr Scorey KC submitted that the Court of Appeal had not explained the juridical nature of a composite policy and whether one is dealing with one policy to which there are multiple parties all subject to the one limit or individual contracts to be read separately or somewhere between the two. However, as Popplewell LJ pointed out the Court of Appeal had decided this issue (in the passage quoted at [73] above): "once it had been decided under issue F that each company was separately insured the answer to this question must be that in general there was a separate limit for each company." Mr Scorey KC sought to maintain, in the face of the Court putting to him that the Court of Appeal in New Hampshire had decided this issue that each insured could be separately insured under a single policy on a multilateral basis as opposed to separate contracts of insurance, that the Court of Appeal had not addressed that issue.
80. In relation to the specific wording of the Starboard policy, Mr Scorey KC drew attention to the fact that in relation to the two extensions to Section 2, Business Interruption on pages 17-18 of the wording (the second of which was cover for Infectious Diseases), it only said Proviso 1 in the Insuring Clause to Section 2 shall not apply to those extensions, which meant that Proviso 2 did apply. That provided: "Provided that…the liability of the Company under this Section shall not exceed: (a) the aggregate Limit of Indemnity as stated in the Schedule". Mr Scorey KC's submission was that the POAND extension (quoted at [14] above) was also subject to that Proviso 2 because of the closing words of that extension: "Subject to the terms, Conditions, limits and Exceptions of this Policy". He accepted that the aggregate Limit of Indemnity was per occurrence (as the judge found, which is not appealed), but the claimants had identified in their pleadings only two occurrences.
81. He also submitted that the Proviso to the POAND extension: "Provided that the Company shall not be liable under this extension for more than the amount shown against this extension in the Schedule" meant that the liability of the insurers was limited to one aggregate limit of £1 million per occurrence. The issue of how many occurrences there were has been held over to a further trial.
82. In relation to the Bath Racecourse wording, Mr Scorey KC submitted that even if "any one loss" remained in proviso (ii) to the DOA clause after the amendment in Condition 22 (the second ground of appeal addressed below), the £2.5 million was an aggregate limit applicable to all the insureds in respect of the same loss. What "any one loss" meant was another issue for the later trial due to take place in May.
83. Mr David Walsh dealt with the two discrete Bath Racecourse grounds of appeal. In relation to the ground about the CPC clause, he accepted that it was common ground that this issue stood or fell with the composite policy issue. He just made two short points. One was that the insurers quibbled at the suggestion in the Bath Racecourse claimants' skeleton that the CPC limit was available to each claimant not just because it was a composite policy but because of the "any one loss" wording. He submitted that the claims preparation costs were insured separately and whether the "any one loss" wording is retained should have no bearing on the cover under the CPC clause. The other point was a related one that the CPC clause is not written on an "any one loss" basis but "any one claim or series of claims arising from a single occurrence.
84. As to why the judge had been wrong to decide that the amended limit for proviso (ii) of the DOA clause in Condition 22 of £2.5 million was still "any one loss", he submitted first that the opening words of Condition 22: "Notwithstanding anything contained herein to the contrary" are not, as the Bath Racecourse claimants assert unremarkable, but unique, not replicated in any of the other Conditions. Mr Walsh submitted that the reasonable policyholder would think that they need not concern themselves with anything else said in the wording about these limits either because anything else would be consistent with what is in the Schedule in which case it adds nothing or if it is inconsistent, Condition 22 prevails. Those opening words of the Condition made clear this was replacement language; £2.5 million replacing £1 million any one loss.
85. Second, he submitted that the amendment to proviso (i) made little sense if the words "any one loss" were intended to be retained because the amount of the existing limit of £1 million was unchanged. This was a further indication that Condition 22 was replacement language. Third, he submitted that it was wrong to approach the limit with an expectation that particular language would be used to achieve a particular end, which was the trap the judge fell into at [369]-[371] when he said that, if the intention had been to delete the "any one loss" language, one would have expected the parties to have said so expressly and clearly. That point cuts both ways, in the sense that if the parties had intended the limit to be retained on the "any one loss" basis, they could equally have used express language to give effect to that intention. Mr Walsh gave the example of the amendment in Condition 17 which made it clear that although the limit was increased it was still "any one occurrence or series of events arising out of one occurrence".
86. Fourth, Mr Walsh submitted that when one reads the conditions as a whole, they do suggest that it is only the lower value limits which are intended to be capable of being accessed on multiple occasions. The conditions which expressly so provide all tend to be in the tens of thousands of pounds, for example condition 6 with a £50,000 limit any one loss. He submitted that in the 41 conditions, the only limits that are for more than £1 million are 22, 23 and 39, none of which contain express language that it is capable of being accessed on multiple occasions. Whilst that could be pure happenstance, he submitted that it was far more likely that the parties objectively intended that the higher limits would not be capable of being accessed on multiple occasions.
87. He also submitted that, on the Bath Racecourse claimants' case, one ended up in the unreal position where the limits on a single business interruption extension potentially ran into tens or even hundreds of millions of pounds, but as I pointed out in argument, that is the case anyway with a £1 million any one loss limit in the original unamended wording, albeit that, as Mr Walsh submitted, the point is starker with a £2.5 million any one loss limit.
88. Mr Kramer KC on behalf of the Bath Racecourse claimants took the lead on behalf of the insureds in responding to the insurers' appeals. He dealt first of all with the overall point whether a composite policy gives rise to a separate contract of insurance for each insured or separate insurance for each within a multilateral contract. Whilst, as he said, it may not matter once one understands that the composite policy gives rise to separate insurances with obligations that are essentially bilateral rather than multilateral, he did point out that the insurers described the composite policy as taking effect by way of separate contracts of insurance in their skeleton argument and that the insurers had not appealed [266] of the judge's judgment which stated:

"It was common ground that the Starboard Policy was a composite policy: that is, a policy which insures the interests of a number of different insured persons in one document, and which took effect legally by way of separate contracts of insurance between Liberty Mutual and each of the individual insured companies."

89. Mr Kramer KC emphasised that the parties agree that the cover under the DOA clause is on a claimant-by-claimant, premises-by-premises basis, so that the claimant needs to point to a danger within one mile of its own premises, preventing or hindering the use of the same premises, rather than some different premises. He submitted that when it came to the limits as amended by Condition 22, the natural way in which the reasonable policyholder would read those was as the limits for their own contract of insurance. They would not think: "these are overall aggregate limits, so I need to find out whether the racecourse in Newcastle or the dog track in Sunderland has already claimed to check the limit is not exhausted".
90. He referred to 1-202 of MacGillivray on Insurance Law headed "Joint and Composite Insurance":

"It has become commonplace for reasons of commercial convenience to insure the interests of a number of insured persons under one policy of insurance, either because it concerns property in which they are all interested, as in General Accident Fire & Life Assurance Corp Ltd v Midland Bank Ltd [1940] 2 KB 388 and State of the Netherlands v Youell [1997] 2 Lloyd's Rep 440; or because they are all companies within one corporate group which can obtain insurance more effectively and cheaply through a single policy…"

91. Mr Kramer KC submitted that a composite policy created separate contracts of insurance with each insured, one consequence of which was that if one insured makes a fraudulent claim, the insurer only has a remedy against that insured; likewise, if one insured commits non-disclosure. Whilst it is possible to contract out of that consequence, unless you do there is a strong presumption that by identifying the policy as composite, the parties intend their risk to be siloed and not affected by the conduct of other insureds. This was well put in argument before Jacobs J in Technip at [136]:

"[that the policy is a composite policy] is important, because – in accordance with the various authorities, including those discussed in Arab Bank plc v Zurich Insurance Co and more recently Corbin & King Ltd v AXA Insurance UK plc HYPERLINK "https://www.bailii.org/ew/cases/EWHC/Comm/2022/409.html" \o "Link to BAILII version" – no conduct or knowledge on the part of one Insured can be attributed to another Insured for the following purposes:

a. Avoidance of the Policy by reason of a non-disclosure or misrepresentation by another insured will not prejudice the claimant insured.

b. The application of an exclusion based on wilful misconduct is limited to the guilty insured.

c. The application of other types of exclusion based on the conduct of an insured other than the claimant insured.

d. The application of a breach of warranty by one insured will not ordinarily affect another Insured.

e. The application of policy limits."

92. He submitted that where there was, as here, a composite policy insuring each insured separately in respect of its premises and business, as a general principle, the limits are also separate for each insured. That can be legitimately described as a presumption or expectation although he accepted in answer to Phillips LJ that this amounted to no more at the end of the day than saying that you are construing a number of separate bilateral relationships, that each insured has contracted separately with the insurers. After a lengthy discussion with the Court, Mr Kramer KC submitted that the fact that this was a composite policy was not neutral as to limits because the reasonable policyholder would naturally, without any contrary specification, understand the limits to be for themselves, not aggregate limits for all the insureds under the policy.
93. He accepted that whether or not limits are individual or aggregate might also depend on the factual matrix, in the sense that the businesses insured might be so interconnected in the sense that they are affected in the same way by profits or losses, making an aggregate limit more likely. Here though, the insureds were entirely separate, miles apart, conducting different businesses with different managers and different financial accounts. Therefore, in a case like the present the Court should proceed on the basis that a composite policy was taken out for convenience, not because the insureds were intending to do anything which intermingled their businesses, although Mr Kramer KC accepted there had been no evidence at trial as to why the insureds took out a composite policy.
94. Mr Kramer KC then dealt with the authorities, starting with New Hampshire. He noted some of the expert evidence before Potter J referred to at 1710; 37-8, setting out the advantages of group or composite policies such as insurers being more interested in underwriting such policies and offering better terms and the administration of premium collection. He referred to what Potter J had said at 1726H-1727A; 49rhc (cited at [75] above) answering the question whether the policy limits applied separately to each group company or there was a single limit applicable to all group companies. He also referred to the judgment of the Court of Appeal already quoted at [73] above that: "once it had been decided under issue F that each company was separately insured the answer to this question must be that in general there was a separate limit for each company." Mr Kramer KC submitted that this was an essential part of the reasoning of the Court of Appeal and binding on this Court, but in any event, it was clearly right.
95. Further submissions on the law were made by Mr William Day. He referred to Colinvaux and Merkin Insurance Contract Law at A-0621:

"Although a composite policy may insure each assured separately in respect of his own loss, it is a matter of construction whether a policy is to be construed as requiring the [insurers] to pay more than the sum insured under the policy where each of the co-assureds has separate claims."

96. He then referred to the three cases which Professor Merkin then cites. It is not necessary to refer to the New Zealand Public Trustee case but the Hong Kong case, of New World Harbourview Hotel Co Ltd v Ace Insurance Ltd [2011] LRLR 230, was one where the judge Reyes J decided that the policy was a composite one and that the HK$100,000 was not a single limit applicable to all claims. The third case cited was Corbin & King which Professor Merkin cites uncritically without any hint of the disquiet caused by that decision in the insurance market to which Mr Scorey KC had alluded but of which, as I pointed out, there was no evidence.
97. In relation to the passages from the judgment of Cockerill J at [230] and [238] cited at [72] above, Mr Day submitted that her reference to "expectation" was another way of referring to the presumed intention of the parties and that presumptions remain important in the construction of insurance contracts. He referred to [163] of the majority judgment of the Supreme Court in the FCA test case where, dealing with proximate causation, it is said: "The requirement of "proximate" causation is based on the presumed intention of the contracting parties…". He also referred to the presumption, outside the insurance context, that neither party intends to abandon any remedies, the so-called Gilbert-Ash principle, recently rearticulated by the Supreme Court in Triple Point Technology, Inc v PTT Public Company Ltd [2021] UKSC 29; [2021] AC 1148.
98. Mr Day submitted that Cockerill J's reference in Corbin & King to expectation was to this type of legal analysis, that it was an ordinary incident of a composite policy to have separate limits, although that starting point can be displaced on the correct construction of the contract. He also referred to Technip. Although that was a case of a composite policy it had an express provision that: "The Policy shall be deemed to be a separate insurance in respect of each Principal Insured hereunder without increasing Underwriters limits of liability" making clear that any limits applied to all insureds, a provision which was completely absent here.
99. Mr Kramer KC summarised the position on this composite policy issue as follows. The ordinary policyholder with their own contract of insurance giving rise to an obligation on the insurer to indemnify them alone, for their loss alone, would expect limits to be applicable to their own insurance and insured interests alone. That is a presumed intention because it is normal unless the facts point otherwise. An insured with their own insurance would not normally agree that their limits were eroded by someone else's claim; likewise, an insured with separate insurance within a composite policy.
100. In relation to the increased limit under Condition 22 of the Bath Racecourse policy, Mr Kramer KC submitted that it was of some relevance that this was not the best drafted policy wording. He submitted that the reasonable reader would understand that, since it was not being highlighted that the basis of the cover had been changed, the basis remained the same and that all that was being changed in proviso (ii) was the increase in the limit to £2.5 million and the addition of the maximum indemnity period.
101. Whilst it was true that the opening words of the Condition: "Notwithstanding anything contained herein to the contrary" were not used in any other Condition, they simply cannot bear the meaning for which the insurers contend. It cannot mean these provisos replace the earlier provisos because that is not what it says. It does not say that it knocks out "any one loss".
102. Mr Walsh's fourth point (referred to at [85] above) did not work, since the original cover of £1 million any one loss was already a large amount, so one cannot read out of the fact that Condition 22 had a large limit an intention to make it a smaller potential limit per insured. Looking at some of the other conditions was useful. For example, Condition 20 expressly stated that the previous wording was "deleted and replaced" and the reasonable policyholder would think that the use of those words was the way to make it clear if "any one loss" were being deleted and replaced.
103. If the intention had been, as the insurers contend, deliberately to move from one basis of cover to another, the reasonable reader would expect that to be spelt out, as in other Conditions, for example the greens Condition 7, which was the point the judge made in [369]. In relation to the insurers' point about proviso (i) which Mr Kramer KC submitted was the only real construction point the insurers have, he submitted the judge was right that arguments based on surplusage are generally weak, especially given that this was not the best drafted wording.
104. Mr Kramer KC did not make any oral submissions about the CPC, noting correctly that the argument on that clause stands or falls with the composite policy point.
105. Mr Gruder KC also made submissions on behalf of the Starboard claimants on the first Ground of the insurers' appeals, which was the only one which concerned them. He submitted that it was not necessary to consider the question of whether there was a presumption or not. Whether the limit was individual or aggregate was one of construction and the framework for the question of construction was that the Starboard policy was a composite policy in one document but with twenty-one separate contracts. So, in relation to a particular insured, one is construing that insured's contract.
106. It was clear that the limit of indemnity in the schedule of £89.9 million referred to in [16] above is a shared limit between all the Starboard insureds, made up of building blocks which are the declared values for each insured. The limits set out in the Schedule under the extensions are effectively arbitrary limits. When one considers the POAND cover (quoted at [15] above]), that is clearly talking about cover for the particular insured and danger or disturbance within a mile of that insured's premises, not someone else's premises. The danger from COVID was different in the case of each insured and its premises.
107. Mr Gruder KC submitted that, accordingly, the proviso: "Provided that the Company shall not be liable under this extension for more than the amount shown against this extension in the Schedule" was referring to the £1 million limit for that particular insured. It made no sense for it to be a shared limit as insurers contend, as it was conceivable that the first claim or the first and second claims would erode the entire limit for the other 19 or 20 insureds.
108. He submitted that this was right not only as a matter of construction, but as a matter of commercial common sense. This was a case of named insureds with identified premises in which each insured carried on its business as hotelier. Each insured did not have any interest in the hotels operated by the other insureds at other premises. If it had been intended that there was to be one shared limit for POAND cover, one would have expected that to be set out clearly in the policy as it would not be what one would expect when considering the POAND clause within the framework of a composite policy. If that had been the intention one would have expected some provision dealing with competing claims or priority of claims and some sort of reinstatement provision.
109. Mr Gruder KC addressed briefly Mr Scorey KC's point about alleged disquiet in the insurance market caused by Corbin & King, emphasising that the effect of writing a composite policy was not a new point but went back at least to Samuel v Dumas and then General Accident and after that New Hampshire, all decided many years before Corbin & King. There was not a shred of evidence of the alleged disquiet and he submitted that only underwriters who were unaware of what they were writing and ignorant of the law could actually have been surprised by the judge's judgment in this case.
110. He submitted that nothing in the "Limit of Indemnity" definition (set out at [18] above) imposed an aggregate limit when read with the POAND clause and the £1 million limit in the Schedule, but the £1 million was the limit under each insured's contract for that insured's loss or series of losses arising from any one occurrence. As Mr Gruder KC submitted, the judge had rejected the insurers' argument that the POAND clause was subject to an aggregate limit and concluded at [246] (cited at [24] above) and at [279] that the only limit in the case of both Gatwick and Starboard was the "any one occurrence" one in the Limit of Indemnity definition. He submitted that he had won on this point and there was no appeal against the judge's finding.

The parties' submissions on the furlough appeals

111. It is common ground that both "Damage" in the Bath Racecourse savings clause (quoted at [13] above) and "Incident" in the Starboard and Gatwick savings clause (quoted at [17] above) are referring to the insured peril.
112. Mr Kramer KC on behalf of the Bath Racecourse claimants cited a passage from the judgment of Derrington and Colvin JJ in the Full Federal Court of Australia in Marrickville at [165] quoted by Butcher J at [98] of Stonegate on the purpose of business interruption insurance: "The purpose of business interruption insurance is to inject additional funds into a going concern to maintain it as a going concern and, in that respect, to return it to an operational state as soon as possible…". He submitted that, if the insurers had acted immediately to fund their insureds their lost revenue, then subject to limits the insured would have had the same profits and cashflow as in a non-Covid year and would have kept its staff and paid their wages. There would have been no need for Bath Racecourse or anyone with such a policy to claim CJRS payments because the insured would have funded its expenses out of normal revenue, which had been replaced. He submitted that crediting the CJRS payments to the insurers is a direct value transfer from the Government to the insurers because, without the payments, the insured would have been in the same position, but the insurers would have paid out more.
113. Mr Kramer KC pointed out that the indemnity under the Bath Racecourse policy, like many business interruption insurances, was not for actual loss at large, but an indemnity calculated by reference to a formula, here on the gross revenue basis. He took the Court through the various items of income referred to in the policy which would count towards the gross revenue. He noted that Government grants, including Covid hardship grants, are income, accounted for as "other income" but they are not income or revenue within the formula in the policy and the insured can keep them without any reduction in the indemnity. Equally, if an insured under this policy wording relied on grant income and it tailed off as a consequence of an insured peril, there would be no cover under the policy for that lost income. However, as Popplewell LJ pointed out in argument, the CJRS grants were not existing grant income, but new grant income after an insured peril.
114. He referred to the provisions in the policy dealing with increased costs of working and other additional costs as a consequence of an insured peril, such as additional interest incurred, and pointed out that the indemnity under the policy also took account of reduced costs due to the peril, which is the savings clause. He said that the examples of "charges" which have ceased or reduced are obvious matters like electricity costs reduced if the insured's factory burns down and submitted that none of the case law or commentary suggests that a savings clause covers a charge which is still incurred but where a source of funding is secured to match it.
115. Mr Kramer KC referred to my decision in Synergy to which both Butcher J in Stonegate and Jacobs J in the present case had referred. He pointed out that it was not a decision about the meaning of "cease or reduce" nor was it obviously a decision about whether an income that sets off a cost means that the cost ceases or is reduced. Depreciation, which that case concerned, is not about income at all. It is just about whether it fell within a "charge" as I held. Mr Kramer KC submitted that the terms of the savings clause should not be strained as a general matter to achieve a full indemnity. It should simply be construed in accordance with ordinary principles of construction.
116. He referred to condition 21 [on p 63] of the wording which excludes certain costs from the analysis whereby there is a netting off of changes in income against changes in costs. Other variable costs reduced would be allowed for in the savings clause.
117. Construing the words of the savings clause in accordance with ordinary principles of construction, "cease" means stop and "reduce" means diminish. Someone stepping in to help pay costs is just not within either of those words and there is no licence for reading the clause by reference to some overall commerciality. Reimbursement or funding is simply not cessation in or reduction of costs. In relation to the example Phillips LJ put to him of the landlord who gives the tenant a rebate or reimbursement in relation to rent already paid, Mr Kramer KC submitted that there the payee was paying something back, whereas here it is not the employee paying back some of their wages but a third party, the Government, funding the reimbursement to the employer, funding the business to incentivise the employer to retain the employees who are not doing any work. Phillips LJ suggested to him that the commercial reality was that the Government paid 80% of the employee's wages, not the employer and saying the insured still paid 80% of the wages ignored that commercial reality. However, Mr Kramer KC maintained the expense of the wages continued to be incurred and neither ceased nor was reduced.
118. Mr Kramer KC referred to Lowick Rose LLP v Swynson Ltd & Anor [2017] UKSC 32; [2018] AC 313, not an insurance case but a professional negligence claim against accountants where the claimant made three loans induced by the accountants' negligence and the borrower was not good for the money. However, for commercial reasons, the owner of the borrower loaned money to the borrower which enabled it to repay two of the loans. The claimant sought to claim the whole amount loaned, contending the repaid loans were funded by the owner and that is collateral or res inter alios acta. The Supreme Court held that the claim failed because the loans had been repaid. Mr Kramer KC referred to [48] in the judgment of Lord Mance JSC who recorded what Longmore LJ had said in the Court of Appeal that "it would be a triumph of form over substance" if a different result occurred "merely because the payment is made through EMSL [the borrower]". Lord Mance explained why it made a difference:

"But the difference is in the nature of the payment, to which Lord Reid referred in Parry v Cleaver [1970] AC 1. Mr Hunt's loan to EMSL was intended to and did lead to actual payment off of the first two loans which Swynson had made to EMSL."

119. Mr Kramer KC submitted that this Court was in danger of falling into the same trap as Longmore LJ had. However, as Phillips LJ pointed out, that was not a case of construction, the issue simply being whether there was a collateral benefit, whereas here there was a question of construction, which was whether this Court was constrained to read "reduce" in the savings clause so as to exclude the situation where a payment is made by the Government, clearly designed to reduce the insured's employment costs.
120. Popplewell LJ also put to Mr Kramer KC that under the savings clause it was the "charges or expenses of the Business" that had to cease or be reduced and the wages bill was a charge on the Business which had been reduced by the CJRS payments. Mr Kramer KC submitted that the charge of the wages by the employee was still incurred. The fact that it was reimbursed or funded by the Government did not change that any more than would be the case if the insured took a loan for exactly the same amount as the relevant charge.
121. Mr Kramer KC referred to the judgment of Osborne J in the Supreme Court of Victoria (Commercial Court) in Princess Theatre Pty Ltd v Ansvar Insurance Limited [2024] VSC 363 another case about JobKeeper payments in Australia. There the savings clause provided:

"Less any sum saved during the Indemnity Period in respect of such of the charges and expenses of the Business payable out of Gross Profit as may cease or be reduced in consequence of the Damage."

The judge noted that at first instance in Marrickville Jagot J had held that credit had to be given for the JobKeeper payments, but the Full Court reached the opposite conclusion. Mr Kramer KC noted that, like the Full Court, the judge in Princess Theatre rejected the insurers' argument based on the general indemnity principle, concluding that the policy provided for an indemnity in accordance with the basis of settlement, in other words with a formula. The judge followed the decision of the Court of Appeal of New South Wales in Mobis Parts Australia Pty Ltd v XL Insurance Company SE (2018) 363 ALR 730 that the indemnity was not against "actual loss" but in accordance with a formula, disapproving of my reasoning in Synergy.

122. The judge concluded at [463]:

"There is no basis for construing the Policy in some way which seeks to give effect to some overriding principle of indemnity. The task is simply one of construing the contract, here the Policy, and determining whether the JobKeeper Payments are to be taken into account in any of the integers of the calculation of the loss of gross profit as set out in the Policy."

Mr Kramer KC submitted that this represented the correct approach to construction, which this Court should adopt.

123. He then turned to the decision and reasoning in Stonegate which Jacobs J followed in the present case in the section of the judgment I have summarised at [41] to [43] above. He accepted that the first of the additional matters Butcher J considered supported his conclusion, accounting treatment, was irrelevant. In relation to the second matter, he submitted that it cannot matter whether the CJRS payment was made before the wages were paid rather than after. The third matter on which Butcher J relied, following Synergy, was that so far as possible the savings clause should be construed: "if there is any room for argument, to accord with the basic principle that the Policy was a contract of indemnity." Mr Kramer KC submitted that to the extent that Butcher J invoked that principle, he was implicitly accepting that some straining of the construction of the savings clause was occurring, but that was wrong. The words "cease or reduce" must be given their natural meaning.
124. Mr Kramer KC referred to the decision of McDonald J in the High Court of Ireland in Hyper Trust Ltd v FBD Insurance plc [2022] IEHC 39. The judge had handed down several judgments, but the case then settled subject to points about the form of order. After the settlement it was conceded that the insured had to give credit for the TWSS payments from the Irish Government. The judgment contains a detailed analysis of whether the requirement to give credit could be justified by reference to the indemnity principle. The judge concluded at [72] that Mobis was consistent with Irish law and should be followed and eschewed reliance on the indemnity principle where the savings clause provides its own formula:

"…the language chosen by the parties must be given appropriate weight. Even where the result may lead either to under-indemnification or over-indemnification, the indemnity principle does not require that a strained or artificial meaning should be given to the words of a policy-at least where there is an evident business purpose for the parties' choice of language… effect is given to the language chosen by the parties even where that does not result in a perfect indemnity. In the present case, the savings clause seems to me to fall into the category of an agreed formula. Accordingly, it is not surprising that its application may potentially give rise to an over-indemnification of the insured. I do not believe that the indemnity principle requires that the savings clause should be interpreted so as to give the words "in consequence of" a different meaning to the way in which those words would ordinarily be understood in an insurance context."

125. Mr Kramer KC submitted that the reasoning of Jacobs J in the present case on this issue was quite short, essentially concluding that Butcher J was right in Stonegate and not dealing separately with the meaning of "cease and reduce". In conclusion on Ground 1 of the furlough appeals, he submitted that the plain words of the savings clause indicate that there was no saving here because the wages were paid in full, did not cease and were not reduced.
126. Mr Kramer KC also made submissions on Ground 3 of the furlough appeals which was that the judge had erred in concluding that credit had to be given for the CJRS payments under the savings clause, because such payments were not "in consequence of" the insured peril but they were collateral/res inter alios acta. He referred to a passage in a textbook, Walmsley on Business Interruption Insurance 2^nd edition at 188 which dealt with the addition of the words "in consequence of the incident" to the standard wording after the decision of Branson J in Polikoff Ltd v North British and Mercantile Insurance (1936) 55 Ll. L. Rep. 279, submitting that these words connoted proximate causation. However, as the Court pointed out, the passage does not say that this is proximate causation, just that the words were introduced to provide some causative link.
127. Mr Kramer KC referred to the decision of the Court of Appeal in Castellain v Preston (1883) 11 QBD 380 and the judgment of Bowen LJ. In that case the insured had contracted to sell the insured property which was damaged by fire. The insurers indemnified the insured against the damage but thereafter the insured recovered the purchase monies from its purchaser. The issue was whether the insured had to give credit for the purchase monies and the Court of Appeal held it did. At 404, Bowen LJ formulated the test as: "can the right to be insisted on be deemed to be one the enforcement of which will diminish the loss?" He then dealt with gifts:

"With regard to gifts, all that is to be considered is, has there been a loss, and what is the loss, and has that loss been in substance reduced by anything that has happened? Now I admit that in the vast majority of cases, it is difficult to conceive a voluntary gift which does reduce the loss."

128. Bowen LJ discussed Burnand v. Rodocanachi (1882) 7 App Cas 333 and then said:

"Suppose that a man who has insured his house has it damaged by fire, and suppose that his brother offers to give him a sum of money to assist him. The effect on the position of the underwriters will depend on the real character of the transaction. Did the brother mean to give the money for the benefit of the insurers as well as for the benefit of the assured? If he did, the insurers, it seems to me, are entitled to the benefit, but if he did not, but only gave it for the benefit of the assured, and not for the benefit of the underwriters, then the gift was not given to reduce the loss, and it falls within Burnand v. Rodocanachi."

129. Mr Kramer KC submitted that one looks at the character of the payment and in the example given by Bowen LJ the benevolence meant that the gift was not something to be taken into account as reducing the loss. He then referred to the decision of Steyn J in Merrett v Capital Indemnity [1991] 1 Lloyd's Rep 169. In that case the issue was whether the plaintiff Lloyd's syndicate which had a claim against its reinsurers had to give credit for some US$45,000 which had been funded ex gratia by the plaintiff's brokers. The judge noted that if the plaintiff's claim for the US$45,000 succeeded it would not receive a windfall, as it had undertaken to account for the sum to the brokers. He said at 170 rhc that:

"On the other hand, if Capitol succeeds they will have received an uncovenanted windfall, in the sense that Merrett will have paid a sum which they were liable to pay, and it seems to me tolerably clear, as the law stands, that the brokers have no claim against Capitol [the reinsurers].

130. At 171 lhc, Steyn J said:

"The payment by the brokers was a gift, albeit a gift made for commercial rather than purely disinterested purposes. The contracts of reinsurance are contracts of indemnity. The question is, therefore, whether the payment diminishes the loss. Not every gift to an assured by a broker diminishes his loss. It is a question of fact in each case whether a gift has or has not been paid in diminution of the loss, and if it is established that the payment was intended solely for the benefit of the assured, it has not been paid in diminution of the loss. In that event it must be disregarded in assessing the assured's recoverable loss…

The arbitrators made primary findings of fact that: (a) the payment was made to retain Merrett's goodwill; that is to benefit Merrett; (b) the brokers expected to be reimbursed by Capitol. It follows inexorably that the payment was made solely for the benefit of the assured and not for the benefit of the reinsurer."

Mr Kramer KC sought to draw an analogy here between the brokers seeking to retain the insured's goodwill and the Government seeking to incentivise the retention of employees for their protection and the benefit of the economy. He submitted that the CJRS payments were nonetheless gratuitous like the funding payment there.

131. He also referred to my decision in Atlasnavios-Navegacao, LDA v Navigators Insurance Company Ltd & Ors ("The B Atlantic") [2014] EWHC 4133 (Comm); [2015] 1 Lloyd's Rep 117. One issue was whether the owners could recover as sue and labour expenses legal expenses incurred in procuring the release of the vessel. Some US$1.2 million of those expenses had been funded by the owners' P&I Club, Gard, on an ex gratia basis, since the case fell outside the scope of Club cover. The insurers argued that the owners could not recover that sum which would represent a windfall, since Gard did not expect reimbursement. I rejected this argument at [348]:

"I agree with [counsel for the owners] that it is a bad point, to which the short answer is that the fact that Gard has funded some of the legal fees is res inter alios acta as between the owners and the insurers. Mr Blackwood QC sought to counter that argument by reference to the brokers' funding cases such as Merrett v Capital Indemnity Corp [1991] I Lloyd's Rep 169, but, as that case demonstrates, where the funding is voluntary (as it was in the present case) and therefore does not diminish the (re)insured's loss, it is to be disregarded in assessing the recoverable loss: see per Steyn J at 171 lhc and MacGillivray on Insurance Law 12^th edition at [34-069]."

132. Mr Kramer KC submitted that the same analysis applied here. The insured was entitled to recover in full without giving credit for the CJRS payments and there was no windfall, since the payments were treated as res inter alios acta since they were voluntary. He also referred to state benefits and local government grants but, in my judgment, they are different from the present scenario and a long way away from it.
133. Mr Kramer KC referred to the analysis of the general law by Butcher J in Stonegate where at [284] the judge said:

"From these cases [which included Burnand v. Rodocanachi, Castellain v Preston and Merrett v Capitol Indemnity], I take the position to be as follows:

(1) If a third party has made a payment which has eliminated or reduced the loss to the insured against which it had insurance, then, subject to the exception below, the insurers are entitled to the benefit of that payment, either in reducing any payment that they might have to make under the policy or, if they have already paid, by claiming the amount from the insured.

(2) This will not be the case, however, if it can be established that the third party, in making the payment, intended to benefit only the insured to the exclusion of the insurers. That might be established if, for example, the third party acted from benevolence towards the insured, as in the case of the brother in Bowen LJ's example in Castellain v Preston; or if that had been expressly stipulated by the third party; or if the third party had paid the money to retain the insured's goodwill and expected to be paid an equivalent amount by the insured's insurer.

(3) In assessing the intentions of the third party payor, it does not matter whether that payor gave any thought to the position of insurers. A payment can still diminish the loss even if no such thought is given."

134. Mr Kramer KC took issue with Butcher J's second point that unless the intention is to benefit the insured to the exclusion of the insurers, the payment falls to be deducted. That seemed to be wrongly putting the burden on the insured to show that intention whereas there was no hint of such a burden in the cases. He submitted that the legal burden was on the insurers but in any event the question was an open one where one just looked to see if the payment was gratuitous etc.
135. He pointed out that in cases such as Merrett and The B Atlantic the donor was well aware of the insurer but said nothing about benefiting the insured alone, without that being a bar to the sum in question being disregarded. One could not read from the silence of the Government that it intended that the insurers should take the benefit of the CJRS payments. Standing back, the general character of the payments was gratuitous, intended to benefit the insured and its employees, not the insurers.
136. Mr Gruder KC addressed the second ground of the furlough appeals on behalf of the insureds, that the CJRS payments were not in consequence of the insured peril because they did not correlate with it. He referred initially to the decision of the Court of Appeal in Bellini v Brit UW Ltd [2024] EWCA Civ 435 which concerned a form of business interruption insurance which only responded where there was physical damage to property. Having cited the well-known recent appellate authorities on construction, Sir Geoffrey Vos MR said at [28] that the relevant clause should be given its natural meaning. Mr Gruder KC submitted that the same approach should be adopted here in construing the savings clause.
137. Mr Gruder KC made a general point that the furlough scheme operated from April 2020 until September 2021 during periods when there was no forced closure of hospitality businesses as well as during periods of lockdown. He emphasised that the savings clause only required sums to be credited to the insurers which were saved in consequence of the insured peril. He submitted, by reference to [162] of the majority judgment of the Supreme Court in the FCA test case that "in consequence of" was wording of proximate causation which meant that the insured peril must be a proximate cause of the saving. The example he gave was of the fire which burnt down a hotel. Savings because the hotel did not have to order any food or pay the electricity bill would be proximately caused by the insured peril, the fire. Here however, the employee is instructed to cease work by reason of the circumstances arising as a result of Covid-19, but that is only the danger part of the insured peril. The insured did not have to show, to receive CJRS, that it had been compulsorily closed.
138. However, as the Court pointed out, the insured peril was a composite one with three elements: (i) danger or disturbance within one mile of the premises; (ii) action by a statutory authority i.e. here the Government restrictions and (iii) preventing or hindering the use of the premises or access thereto (in the case of all the policies under consideration but in the case of the Starboard and Gatwick policies also, as an alternative, interference with the business of the insured). In other words, in the case of the Starboard and Gatwick policies, the third element did not require prevention or hindrance of use or access but was satisfied by interference with the business. As I pointed out in argument, even when the first lockdown ended, there was interference with the business of companies running hotels because of social distancing requirements and the other restrictions, certainly in the case of Gatwick, whose policies ran to 20 October 2020, even if Starboard's policy expired on 30 June 2020, before the first lockdown was lifted.
139. Mr Gruder KC maintained that the savings through CJRS payments were not in consequence of the insured peril because a business did not have to prove any element of the insured peril to claim CJRS under the furlough scheme. The Court challenged that contention in argument, putting to him that each of the three elements of the insured peril was a cause of the lockdown and thus of the furlough scheme. Mr Gruder KC contended that this was not the case and submitted that the insurers could not show that payment under the CJRS was in consequence of being prevented or hindered by Government/statutory authority by reason of a case or cases of Covid within one mile of each of the premises.
140. Mr Gruder KC referred to [2.1] and [6.1] of the Treasury Direction quoted at [9] above and pointed out that a business might have had a general downturn because of the pandemic, but not been ordered to close and it would still be entitled to claim under the CJRS. To be entitled to claim, there is no requirement that there be Covid within one mile of the premises or that the business has been instructed to close by the Government. All that is required is that the instruction to furlough employees is given by reason of circumstances arising as a result of Covid.
141. He submitted that Stonegate was of no assistance since in that case it was conceded by the insured that the CJRS payments were received as a consequence of the insured peril and in any event the insured peril was disease in the Vicinity, the definition of which encompassed the whole of the UK. There was no requirement that Covid be within a particular distance of the premises or of an instruction from the Government. Butcher J was focusing on the disease peril in that case.
142. He referred to how the issue whether the cesser or reduction of charges and expenses was "in consequence of the interruption or interference" was dealt with by the Full Federal Court in Marrickville at [461]:

"The requirement is that the cessation or reduction of charges or expenses be in consequence of "the" interruption or interference. This takes the reader back to the beginning of the "Cover" section, where reference is made to the insured's business being interrupted or interfered with "as a result of Damage". It is that interruption or interference that is being referred to in the "sum saved" provision. The word "Damage" is given an extended meaning by cl 8. Relevantly for present purposes, the "outbreak of a human infectious or contagious disease occurring within a 20-kilometre radius of the Situation" is deemed to be Damage to Property used by the insured at the Situation. Thus, the reference (at the beginning of the "Cover" section) to the insured's business being interrupted or interfered with "as a result of Damage" is to be read as including the insured's business being interrupted or interfered with as a result of the insured peril described in cl 8(c). Returning, then, to the "sum saved" provision, the concern is with "the" interruption or interference earlier identified, that is, the interruption or interference resulting from the insured peril in cl 8(c). The question, then, is as follows: assuming that Meridian is able to establish on evidence that the insured peril in cl 8(c) is a proximate cause of its loss (as to which, see PJ [481], [485] – [498]), were the JobKeeper payments made and received "in consequence of" the interruption or interference (that is, the interruption or interference resulting from the insured peril in cl 8(c))? As a matter of the application of the policy's provisions, they were not. The criteria for eligibility for JobKeeper payments were financial ones; they did not depend on whether or not there had been an outbreak within 20 km of the premises of the business. Meridian was entitled to the JobKeeper payments regardless of whether or not there was an outbreak within 20 km of its premises. Conversely, had Meridian not met the financial tests for JobKeeper, it would not have been entitled to JobKeeper payments, even if the insured peril in cl 8(c) occurred. Accordingly, the second aspect of the "sum saved" provision (the causal requirement) is not satisfied."

143. Mr Gruder KC pointed out that the Full Federal Court was saying that whilst disease within 20 kilometres of the premises entitled the insured to an indemnity, that was not a requirement for the JobKeeper payments which were based on a totally different financial criterion. He submitted that the position was exactly the same under the CJRS.
144. In relation to the judgment of Jacobs J in the present case, Mr Gruder KC submitted that when the judge said at [443] that the CJRS and the restrictions introduced because of the prevalence of Covid-19 were "closely connected" that was not the same as proximate causation and this was where the judge went down the wrong path. He submitted that the judge's assessment at [445] that the same approach should be adopted to causation in relation to the savings clause as in relation to the composite insured peril was wrong. There were other businesses which could claim CJRS payments merely by showing that their business had been affected by the pandemic, without having to prove the three elements of the insured peril. He submitted that the proximate cause of the CJRS was not the insured peril, since Covid within one mile of the premises and prevention of access or hindrance of use were not relevant to receipt of CJRS payments under the scheme.
145. On behalf of the insurers, Mr Scorey KC submitted that, although Mr Kramer KC had identified three issues, cessation/reduction, causation and collaterality, there were really only the first two, since collaterality is not really separate from causation. He submitted that, in construing the wording, it was important to bear in mind the indemnity principle. The insureds' submissions were to the effect that the indemnity principle should be ignored, as one is not concerned with actual loss but just an agreed mechanism. However, these were not agreed valuation policies. The main component of the business interruption loss was reduction in revenue, but as a measure of loss that only makes sense if one takes into account the saved operating expenses.
146. He drew attention to the definition of "Standard Gross Revenue" in the Starboard and Gatwick policy wordings under which the figures adjusted: "shall represent as nearly as may be reasonably practicable the results which
     but for the Incident would have been obtained", in other words seeking to identify the actual loss. Similar language is found in the Bath Racecourse wording in the definition of "Standard Gross Revenue" and the trends clause. He submitted that the reasonable policyholder reading these policies would assume this is seeking to identify the loss in terms of the economic burden actually sustained.
147. Construction was also to be looked at from the perspective of the reasonable policyholder, who would expect the words "in consequence of" used in different places in the wording to mean the same thing. If they mean one thing in terms of the insured peril, they should mean the same elsewhere, here in the savings clauses.
148. Mr Scorey KC submitted that Bellini, to which Mr Gruder KC had referred, is irrelevant. It concerned an attempt by the policyholders to correct the wording of the contract which they alleged was mistaken or had gone awry. That is not an issue in this appeal, which simply seeks to determine what the terms of the savings clauses mean. That is clearly correct and, in my judgment, Bellini is of no particular assistance in the present case.
149. Mr Scorey KC submitted that the insureds' submissions on the cessation/reduction issue embraced form rather than substance. The substance was that under the CJRS, if the employer continued to pay the furloughed employees their wages, the Government would refund 80% of the monies via the PAYE scheme. The economic effect of the CJRS was that the employer did not have to bear the economic burden of the wages bill. A reasonable policyholder, if asked the question: "if you are given this 80% grant under the CJRS, is that a saving?" would say: "of course, it is not an expense borne by my business". Butcher J in Stonegate and Jacobs J in the present case were right to look at the economic commercial reality, which is the correct approach to the construction of commercial contracts.
150. This issue of cessation or reduction was not addressed by the Full Federal Court in Marrickville at all, which was only concerned with causation. The suggestion in the Bath Racecourse claimants' skeleton that Jagot J at first instance was sceptical about the reduction point was wrong. Jagot J said at [623]-[624]:

"623 Further, the amount saved provision is also potentially applicable. If the amount paid by a third party means that an expense of the business is reduced, then that reduction must be taken to be in consequence of the interruption or interference unless the payment is incapable of being so characterised. Again, an insured cannot have it both ways. If the cover extends to the circumstances underlying the insured peril, then the expenses saved must also extend to the circumstances underlying the insured peril.

624 For the reasons already given, JobKeeper payments reduce the insured's loss and expenses in the form of saved wages' payments…"

The judge therefore took an economic view, very much akin to that of Butcher J and Jacobs J.

151. In relation to the causation issue, Mr Scorey KC noted that it was common ground that the "incident" or "damage" in the savings clauses was referring to the insured peril and that the words "in consequence of" import a causation analysis. He noted that Jacobs J had correctly found at [445] that "in consequence of" does not necessarily mean a strict and narrow causation analysis where one had to track precisely the insured peril for this policyholder as the sole and proximate cause of the furlough scheme.
152. He submitted that whether one was looking at causation for the issue of whether the claim is covered or for the trends clause or the savings clause, the test is whether each element of the insured peril was a proximate cause of the relevant saving. If that test was satisfied, the precise terms of the CJRS and the entitlement to payment are irrelevant.
153. Mr Scorey KC submitted that the incidence of Covid within the relevant radius was a sufficient effective cause of the restrictions imposed and must be equally a sufficient effective cause of the furlough scheme introduced at the same time as the restrictions to mitigate their effect. This was entirely orthodox in light of the approach of the Supreme Court in the FCA test case.
154. In the case of the Bath Racecourse policy, the starting point is that the claimants' case is that they can satisfy the insured peril, that action by the Government following danger or disturbance within a one-mile radius of the relevant premises has prevented or hindered use of the premises or access thereto. It is because of that prevention or hindrance in consequence of the restrictions that the furlough scheme was brought in to mitigate their effect. If there had not been the prevention or hindrance, the furlough scheme would not have been introduced. As Popplewell LJ put it in argument, the fact that the prevention or hindrance may have involved other premises not insured is irrelevant because of the proximate cause analysis and rejection of but for causation in the FCA test case.
155. This was a statutory scheme announced at the same time as the restrictions were introduced to mitigate their effect and gave rise to a public law right of employers who furloughed their employees to claim furlough payments corresponding to 80% of the wages. It was not a gift or donation. It was nothing to the point that employers who furloughed their employees included those whose use of or access to their premises was not prevented or hindered. The policyholders' argument ignored the causation test of the Supreme Court in the FCA test case under which it was the cases of Covid within the relevant radius of these policyholders' premises which was the proximate cause of the restrictions and, therefore, of the furlough scheme and the fact that there were other areas of the country not affected is irrelevant. The example Mr Gruder KC gave of the corner shop which did not close but laid off some of its employees was not relevant, because the owners of that hypothetical shop were not making a claim under this insurance.
156. Mr Scorey KC submitted that logically, if a single case within the radius is good enough to establish the insured peril and concurrent causation for the purposes of the imposition of the restrictions, it must follow that is good enough to establish causation for the introduction of the CJRS, not least because it is artificial to dissect the CJRS as being totally separate from the restrictions. They go hand in hand. He endorsed the approach of the judge at [446] that what works one side of the line should work the other, which is supported by the Supreme Court which adopted the same approach to the trends clauses as to the insured peril.
157. In relation to Marrickville Mr Scorey KC noted, as had Jacobs J, that the Australian JobKeeper scheme was a different scheme and, in any event, the points now being ventilated about the savings clause mirroring the three-stage composite peril do not seem to have been ventilated before the Full Court. If necessary, he submitted that Marrickville was wrongly decided, at least as a matter of English law, and should not be followed.
158. In relation to collaterality, he had already submitted that this was intimately intertwined with causation. He submitted that there was a danger in trying to be too clever in this area, trying to find a unifying factor in the common law applicable in the spheres of tort, contract and insurance. This case concerns the wording in an insurance policy to be understood and applied by a reasonable policyholder and one should not invite extraneous considerations into the analysis, such as public policy considerations in the law of tort. The CJRS payments were neither collateral nor benevolent gifts for the reasons correctly given by Butcher J in Stonegate and followed by Jacobs J in the present case. The passage at [284] of Butcher J's judgment quoted at [132] above is an accurate summary of the legal principles. The starting point is that if a third party has made a payment to the insured prima facie that goes to reduce the loss unless it falls within various exceptions, such as that at [284(2)], that the third party intended to benefit only the insured and not the insurers. The CJRS payments do not fall within any of those exceptions.
159. In relation to state benefits, it was now clear from Hodgson v Trapp [1989] AC 807 that prima facie they will reduce the claimant's loss (see per Lord Bridge of Harwich at 822D-E). Mr Scorey KC submitted that the CJRS payments were pursuant to a statutory scheme funded by taxation and, in the same way as state benefits, should reduce the insured's loss under the policy. The contrary conclusion could only be reached if the intention of the Government had been that the insurers should not benefit from the CJRS payments, but there was no indication of such an intention.
160. He submitted that Merrett and The B Atlantic were cases of ex gratia payments for sound commercial reasons to generate or retain the goodwill of the payee and so those payments were not taken into account, but those cases were nothing to do with a scheme such as the present.

Discussion

161. The first Ground of the insurers' appeals concerns whether, under a composite policy such as the Bath Racecourse and Starboard policies, the limits for the respective DOA and POAND clauses in those policies were shared aggregate limits applicable to all claims by the insureds as a whole or individual limits for each insured. The starting point, as I see it, is that the composite policy contains in one document what is, as a matter of legal analysis, a series of separate contracts of insurance insuring each of the insureds. As Mr Kramer KC pointed out, the insurers have not appealed the judge's finding to that effect at [266] which also records that, at least as regards Starboard, this was common ground at trial. In any event, the analysis that a composite policy comprises a series of contracts of insurance with each policyholder insured separately is supported by the authorities.
162. Thus, in New Hampshire, Potter J, having discussed the question whether the policy was joint or composite, concluded that it was composite and said at 1715H; 147 rhc: "Accordingly, I am satisfied that in this case the insured companies were insured severally in respect of their several interests". That conclusion was confirmed by the Court of Appeal in the passage quoted at [78] above. The submission by Mr Scorey KC that the Court of Appeal had not grappled with that issue is unsustainable and flies in the face of what the Court actually said in that passage. The analysis that there is a series of contracts of insurance insuring each policyholder separately also emerges from Corbin & King, where, at [231] Cockerill J said:

"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."

163. Since there is no evidence whatsoever to support the alleged disquiet in the insurance market caused by Corbin & King, it is not necessary to dwell on that point, save to say that there was considerable force in Mr Gruder KC's submission that the law on the effect of composite policies goes back a long time to Samuel v Dumas and then General Accident and, if insurance market practitioners were surprised by the judgments in Corbin & King or in the present case, it can only have been through ignorance of the law.
164. Given that the composite policy comprises a series of contracts of insurance insuring each policyholder separately, it is clear (and indeed was not disputed by the insurers) that the DOA clause in the Bath Racecourse policy and the POAND clause in the Starboard policy are to be construed as applying to each insured policyholder separately, in other words the "Business" with which there has been interruption or interference is the business of that insured and no other and the "Premises" affected are those of that insured and no other. In the case of both the Bath Racecourse claimants and the Starboard claimants, each insured only had an interest in its own racecourse etc or hotel and did not have any interest in premises operated by the other insureds. Furthermore, the danger from Covid-19 was different in the case of each insured and its premises. This is demonstrated by the fact that, as I pointed out in argument, one of the Starboard insureds was the owner of the Ibis in Leicester, which was a Covid hotspot, so the danger was more extreme for that insured than for a Starboard insured which owned a hotel elsewhere in the country.
165. I agree with Mr Gruder KC that, in those circumstances, if it had been intended that there was only one shared aggregate limit for the DOA or POAND cover, one would have expected that to be clearly set out in the policy wording, together with some provision to deal with priority of competing claims. In my judgment, in the absence of any provision in the insurance that the limit in the DOA clause in the Bath Racecourse policy or the POAND clause and Extension 15 in the Schedule to the Starboard policy was intended to be an aggregate one (as was the case in Technip), the correct construction of those limits is that they are applicable to each insured separately.
166. I do not consider that it is necessary to rely upon any "presumption" that there are separate limits applicable to each insured in the case of these composite policies. Rather, the conclusion that the limits are ones applicable to each insured and not aggregate limits is a matter of the proper construction of the respective policies. That approach accords with what is said by Professor Merkin in the passage cited at [95] above. I also agree with Mr Kramer KC that it is what the reasonable policyholder under these composite policies would expect to be the position with regard to the limits. In the absence of some provision, as in Technip, that the limits were in the aggregate across all the insureds, the reasonable policyholder would not expect its own limits to be eroded by someone else's claim.
167. In the case of the Bath Racecourse policy, the words "any one loss" in proviso (ii) to the DOA clause are referable to the loss of the particular insured under consideration, not to an aggregate limit, as each insured has suffered its own loss in respect of its business at its premises. Similarly, the words in the POAND clause in the Starboard policy: "Provided that the Company shall not be liable under this extension for more than the amount shown against this extension in the Schedule" were referring to the £1 million limit for each particular insured. Likewise, the definition of "Limit of Indemnity" in the Starboard policy: "for any loss or series of losses arising from any one occurrence" is pointing to the loss suffered by the particular insured, even if there are only a limited number of occurrences affecting all the insureds. The judge concluded at [246] that the limit set out in the "Limit of Indemnity" provision is an "any one occurrence" limit and there is no aggregate limit in that provision or the POAND clause. As Mr Gruder KC pointed out, there is no appeal against that finding.
168. It was accepted by the insurers that, if the Court reached the conclusion that the limits were separate limits for each insured and not aggregate limits, the same conclusion should be reached in relation to the CPC clause, so it is not necessary to say any more about it.
169. So far as Ground 2 of the insurers' appeals is concerned, I have reached the firm conclusion that Condition 22 in the Bath Racecourse policy increased the limit under the DOA clause proviso (ii) to £2.5 million any one loss from £1 million any one loss and did not delete or remove the "any one loss" wording from the proviso. I have reached that conclusion for a number of reasons. First, although, as was effectively accepted by both parties, this policy wording is not a model of clarity or grammar, I consider that if the intention had been that the basis of cover was being changed so fundamentally from "any one loss" applicable separately to each loss suffered by each insured to an aggregate limit applicable to all insureds, the wording would have made that change clear, for example by including words to the effect that "any one loss" was "deleted and replaced" as in the case of Condition 20 or "deleted" as in the case of Condition 34.
170. Second, and closely related to the first reason, if it had been intended that the basis of cover would change to being in the aggregate, one would expect the Condition to have made this clear by express words such as "in the aggregate" as in Condition 7, the damage to greens provision.
171. Third, I agree with Mr Kramer KC that, whatever is meant by the opening words of the Condition: "Notwithstanding anything contained herein to the contrary", they cannot bear the meaning for which the insurers contend, that they delete the "any one loss" wording. The argument of the insurers is essentially a circular one, because it assumes in their favour that there is something in the existing wording which is contrary to the Condition, which, in the absence of express words in the Condition stating that "any one loss" is deleted and replaced, there is not.
172. Fourth, I also agree with Mr Kramer KC that there is nothing in the insurers' suggestion that the limit of £2.5 million must be an aggregate limit because it is such a large limit compared with other limits in the Conditions. The original unamended cover had what on any view was a large limit of £1 million any one loss compared with those other limits and it is not possible to read out of it being increased to £2.5 million an intention to make it a smaller limit per insured.
173. Fifth, I do not consider that there is any force in the insurers' argument based on proviso (i) that the amendment in Condition 22 made no sense if the words "any one loss" were intended to be retained, because there was no need to refer to "GBP 1,000,000" as being an amendment, given that that was the limit in the unamended DOA clause. Mr Kramer KC was correct that arguments based on surplusage are generally weak, especially given that this was not the best drafted wording. As Lord Leggatt JSC said in Triple Point at [119]: "arguments of this sort based on verbal surplusage in a commercial contract do not count for much."
174. Turning to the insureds' furlough appeals, Ground 1 raises the argument that the insureds' wages costs did not "cease" and were not "reduced" by the CJRS payments. In my judgment, the insureds' argument does, as Mr Scorey KC put it, embrace form over substance. Looking at the substance, the commercial and economic reality by reference to which these commercial insurance contracts should be construed, is that the effect of the CJRS payments reimbursing the insureds for 80% of their wages bill was indeed to reduce that bill by 80%, which constituted a saving within the savings clauses. It is no answer to that commercial and economic reality, as the insureds seek to contend, that the insureds still paid the wages and that the funds reimbursing the insureds came from a third party, the Government. The bottom line at the end of the day is that the insureds did not have to bear the expense of the wages bill and to that extent, the charges or expenses of the business were reduced. I agree with Mr Scorey KC that this is how the reasonable policyholder would view the position.
175. It follows that on this issue of cessation or reduction I agree with what Butcher J said at [258] of Stonegate (followed and applied by the judge in this case) which merits repetition:

"In my judgment, employment costs were at least 'reduced' pro tanto by reason of the payment of corresponding amounts under the CJRS. I consider that the natural meaning of the definition, including its savings clause, is that it is referring to costs to the business. Insofar as such costs were defrayed by the government, I consider that they were 'reduced'. That, in my view, reflects the net financial effect of payments under the CJRS and the commercial reality."

176. There is nothing in the Australian case of Marrickville which would support a contrary conclusion. As Mr Scorey KC correctly pointed out, the judge at first instance, Jagot J, found that the JobKeeper payments did reduce the costs of the insured's business in the passages quoted at [149] above, taking an economic view similar to that of Butcher J and Jacobs J. The issue was not before the Full Federal Court on appeal. Osborne J in the Supreme Court of Victoria in Princess Theatre reached the opposite conclusion at [466] that the wages did not cease and were not reduced by JobKeeper payments. In my judgment, that conclusion does not represent English law in respect of these savings clauses, which has been correctly stated by Butcher J and Jacobs J.
177. I have reached the conclusion that the charges or expenses of the insured's businesses were reduced by the CJRS payments as a matter of commercial and economic reality, irrespective of the three additional matters upon which Butcher J relied as supporting his conclusion. So far as those three matters are concerned, I consider that how the payments were treated as a matter of accounting by the insureds is not relevant and that the question whether the costs or expenses were reduced cannot depend upon the timing of the CJRS payments.
178. However, I do agree with Butcher J that the savings clauses should be construed to accord with the basic principle that the policies are contracts of indemnity. As noted above, much was made by Mr Kramer KC, in his submissions on this ground of appeal, of a point that the indemnity under these business interruption insurances was not for actual loss but for reduction in gross revenue in accordance with a formula which, as he pointed out in relation to the Bath Racecourse policy, excluded under Condition 20 items such as hospitality costs, prize money, raceday savings and television rights income. However, in my judgment, this argument misses the point. The indemnity may not be in respect of the total loss (and the exclusion of the items in Condition 20 may well be because, for example, they may not be incurred or incurred in full if the DOA clause bites and so might fall within the savings clause in any event), but the clear intention is still, after adjustments in accordance with the formula have been made, including for the trend of the business, that the adjusted figures for gross revenue "shall represent as nearly as may be reasonably practicable the results which but for the [Damage/Incident] would have been obtained during the relative period after the [Damage/Incident]", words in the definitions of Standard Gross Revenue in all the policies. Those words demonstrate clearly, as Mr Scorey KC submitted, that the policy is seeking to identify the actual loss suffered by the insured, since it is the amount by which the gross revenue during the indemnity period falls short of that standard gross revenue as defined which the insurers agree to pay as indemnity, subject of course to applicable limits.
179. Whether my actual decision in Synergy was right or wrong (and I am aware it has not been followed in Australia), the principle which Butcher J derived from it holds good, namely that "as a matter of principle, a policy should be interpreted as providing an indemnity for the loss suffered not for more than such an indemnity", unless the wording of the policy dictates a different result, which none of the policies does in the present case. In my judgment, the application by Butcher J of that principle in Stonegate at [269] is equally applicable in the present case:

"The precise issue in Synergy Health was different from that here, in that it concerned a saving of depreciation. Nevertheless, I consider that the approach in that case is one which I should adopt in this case as well. The CJRS payments were in respect of an expense of the business, and resulted, in reality, in a saving of cost. For the clause to be construed so as to mean that those payments were not counted as savings would, in my view, mean that the insured would receive more than an indemnity. It should, if possible – and in my view it clearly is possible – be construed so that those payments are taken into account under the savings clause."

180. Ground 2 of the furlough appeals raises the question of causation, whether the cessation or reduction of the charges or expenses of the business was "in consequence of" the Damage/Incident, it being common ground that these words import a causation analysis and that the "Damage" or "Incident" is a reference to the insured peril.
181. The effect of the decision of the Supreme Court in the FCA test case is that, in the case of DOA or POAND clauses such as are under consideration here, the insured peril is a composite one with three elements: (1) danger or disturbance within a one-mile radius of the insured's premises (which here was the occurrence of at least one case of Covid-19 within that radius) which causes (2) action by a statutory authority (here the restrictions imposed by the Government) which cause (3) prevention or hindrance in the use of the insured's premises or access thereto, or alternatively, in the case of Starboard and Gatwick, interference with the business of the insured: see [215], [243] and [250] of the majority judgment. In assessing whether the insureds had suffered losses in consequence of the insured peril, the Supreme Court adopted a concurrent causation analysis, that it was enough if the Government action was taken in response to cases of Covid-19 which included at least one such case within the relevant geographical radius, even though the other cases which occurred outside the radius were uninsured, provided that they were not excluded under the policies, which they were not: see [173]-[176] of the majority judgment. In doing so, they rejected a but for test of causation, both in relation to the operation of the insured perils and in respect of the trends clauses.
182. In my judgment, the starting point on this issue of causation is to consider what is the purpose of the savings clause. It is to ensure that there is deducted from the calculation of the reduction in gross revenue sums saved because the charges or expenses of the business have ceased or been reduced as a consequence of the insured peril. In principle, since this is looking at reduction of the loss as a consequence of the insured peril, one would expect that the same approach should be adopted as in the assessment of whether the loss claimed by the insured is in consequence of the insured peril. The savings clauses, like the trends clauses considered by the Supreme Court, are part of the machinery contained in the policies for quantifying the loss and, as with those trends clauses, should be construed consistently with the insuring clauses in the policies: see [260]-[262] of the majority judgment.
183. As the judge put it at [445]-[446], which bears repetition:

"However, the effect of the decision of the Supreme Court in the FCA test case is that, when considering the operation of the insured peril, a concurrent causation analysis is to be applied. It is therefore sufficient, for the purposes of coverage, for a policyholder to show loss flowing from a combination of an insured peril which affected its business together with similar perils which affected other businesses. I consider that the same approach can and should properly be taken when considering causation in the context of the receipt of CJRS payments. It is therefore sufficient to show that the CJRS (and thus the payments made pursuant to that scheme) was brought into being in consequence of a combination of government restrictions affecting the business of each claimant policyholder in combination with restrictions affecting the business of other policyholders.

…Thus, as Mr Scorey submitted, the case against the insurers in relation to the peril is a concurrent causation analysis: there was a relevant action by the statutory authority following disease within 1 mile of the premises, and that interfered with the policyholders' business. Equally, the furlough savings were in consequence of what had happened: they were brought in because of damage to businesses caused by the restrictions on a large number of businesses, including those of the claimants, brought in by the government as a result of the pandemic. I agree with Mr Scorey that what works on one side of the line should also work on the other, and that it is not appropriate to take a different and much stricter approach to causation in the context of savings than in the context of the insured peril."

184. Mr Gruder KC's principal argument as to why the savings through CJRS payments were not in consequence of the insured peril was that a business did not have to prove any element of the insured peril to claim CJRS payments under the furlough scheme. Hence his example of the corner shop which did not close during the pandemic, but which furloughed some of its employees because of a general downturn in the business. It would not satisfy the insured peril, but it would satisfy the criteria for CJRS payments under the Treasury Direction. The short answer to that example is that that business is not making a claim under this insurance. The question here is whether the insureds, each of whom can establish the three elements of the insured peril (otherwise they would not have a claim under the insurance at all), have to give credit for the CJRS payments they received against the loss they claim under the policy. It is irrelevant to that question that there are other businesses not insured who could claim under the furlough scheme, just as the fact that a prevention or hindrance may have affected other premises not insured is irrelevant pursuant to the causation analysis of the Supreme Court. Likewise, it is irrelevant that there were periods of time when businesses claimed furlough notwithstanding that there was no forced closure of the business, between lockdowns. The issue on these appeals is whether the insureds have to give credit for CJRS payments against insured losses, not against uninsured losses.
185. As the Court put to Mr Gruder KC, taking each of the three elements of the insured peril in turn, it is clear that the CJRS payments were in consequence of each of those elements and that the test of causation under these savings clauses is satisfied. Thus, taking the first two elements together, it was the general prevalence of Covid-19 (including cases within the relevant radius) which led to the restrictions imposed by the Government. The furlough scheme was announced at the same time that those restrictions were imposed and was intended to mitigate the effects of those restrictions, so that the incidence of Covid-19 and the restrictions imposed as a consequence were a sufficient effective cause of the furlough scheme. As Mr Scorey KC said, this is entirely orthodox in the light of the approach of the Supreme Court in the FCA test case.
186. In relation to the third element of the insured peril: prevention or hindrance in the use of the insured's premises or access thereto (or, in the case of Starboard and Gatwick, interference with the business) that flowed from the restrictions imposed by the Government was something that affected all the insureds; otherwise they would not have had claims under the policies. Likewise, it is clear that that prevention or hindrance or interference caused the insureds to have to furlough employees, which in turn led to their claiming CJRS payments.
187. The arguments for the insureds on the savings clause focused on the terms of the Treasury Direction and what was required for a business to claim a CJRS payment, but in my judgment, the correct focus should be on whether the reduction in the charges and expenses of the businesses of these insureds was in consequence of the insured peril (which for the reasons set out above it clearly was), not on the general question what criteria needed to be satisfied to claim under the CJRS. The insureds' arguments in effect depend upon a but for test of causation: it is said that the insureds would have been entitled to CJRS payments in circumstances in which the insured perils had not occurred and so would be entitled to such payments, without the occurrence of the insured perils, and so the savings were not caused by the insured perils. But these are the very but for causation arguments which were rejected by the Supreme Court in the FCA test case, both in relation to the operation of the insured perils, and the trends clauses. They are equally inapposite to the related causation question which applies to the savings clauses, in which the proximate cause test should be the same. For the same reason, at least so far as the correct construction of these policies as a matter of English law is concerned, the approach of the Full Federal Court in Marrickville in the passage cited at [141] above suffers from the same error of focusing on the criteria for satisfying the JobKeeper payments scheme, rather than the question whether the savings made were in consequence of the insured peril. As Mr Scorey KC said, the Full Federal Court does not seem to have had ventilated in the argument before it the critical point here that the savings clause should mirror the insured peril.
188. Ground 3 of the furlough appeals is that payments made under the CJRS were not "in consequence of" the insured peril because such payments were collateral or res inter alios acta. Despite the ingenuity of the arguments of Mr Kramer KC on this issue, they are misconceived. As Mr Scorey KC submitted, the question of collaterality is inextricably connected with that of causation and once it has been determined, in relation to Ground 2, that the reduction in the charges and expenses of the business is in consequence of the insured peril, the argument that the payments which led to that reduction were collateral must fail.
189. The CJRS scheme was a statutory scheme designed to mitigate the effects of the restrictions imposed by the Government during the pandemic and gave businesses which satisfied the criteria under the Treasury Direction what was in effect a public law right to receive CJRS payments. There is no analogy whatsoever with benevolent gifts or with ex gratia payments, even if made for commercial reasons, as in Merrett v Capitol Indemnity or The B Atlantic. A more apt analogy is with state benefits which, as Hodgson v Trapp confirmed, will prima facie reduce the claimant's loss. Mr Kramer KC submitted that state benefits could only do so if part of a pre-existing benefit system rather than a response to insured events. However the decision in Randal v Cockran (1748) 1 Ves. Sen. 98, 27 ER 916, which was approved but distinguished in Burnand v Rodocanachi, shows that benefits in specific response to a catastrophe may do so. In that case the victims of attacks by Spanish ships were, by Royal proclamation, granted a share of prizes from goods seized in reprisals, and it was held the benefits went to reduce the liability of insurers.
190. In my judgment, the passage in the judgment of Butcher J in Stonegate at [284] quoted at [132] above correctly summarises the law in this area. A payment by a third party which reduces the loss of an insured who has insurance (as in the present case) will reduce the amount for which the insured can claim under the policy, unless it can be established that the third party, in making the payment, intended to benefit only the insured to the exclusion of the insurers. Examples where that exception applied are Burnand v Rodocanachi, where the United States had made it clear that it was not paying for the purpose of reducing the loss against which the insurers had indemnified, but for a different purpose and Merrett v Capitol Indemnity where, as Steyn J said: "It follows inexorably that the payment was made solely for the benefit of the assured and not for the benefit of the reinsurer."
191. However, in the present case, there is no question of that exception applying, as Butcher J made clear at [286] of Stonegate:

"As to the intention of the Government in paying, Stonegate has not shown that this was with the intention of benefiting Stonegate alone to the exclusion of insurers. There is no express statement by the Government to that effect. The Government did not indicate that the payment was being made only in respect of uninsured losses. This is notwithstanding that, unsurprisingly, the Government was aware that some companies had BI insurance, as evidenced by the Treasury's Fact Sheet of 18 March 2020."

The CJRS payments were neither collateral nor res inter alios acta and the insureds are obliged to give credit for them under the savings clause.

Conclusion

192. For all the reasons set out above, I consider that both the insurers' appeals and the insureds' furlough appeals should be dismissed.

Lord Justice Popplewell

193. I agree.

Lord Justice Phillips

194. I also agree.