Insurance contract law: causation

PRACTICE NOTE

Reproduced from Practical Law, with the permission of the publishers. For further information visit uk.practicallaw.thomsonreuters.com

byAndrew Fletcher KC and Adam Kramer KC, 3 Verulam Buildings and Practical Law Financial Institutions

Status: Law stated as of 17 June 2023 | Jurisdiction: England, Wales

This document is published by Practical Law and can be found at: uk.practicallaw.tr.com/w-039-3868

Request a free trial and demonstration at: uk.practicallaw.tr.com/about/freetrial

An overview of the law of causation in general insurance contract law. The note explains the

principle of proximate cause, the ﬁndings of the Supreme Court in Financial Conduct Authority v

Arch Insurance (UK) Ltd and others [2021] UKSC 1 relating to causation where there are multiple

concurrent causes of loss and the relevance of causation when calculating the insurance indemnity.

Scope of this note

The concept of causation is central to many insurance

claims, most especially as it regulates the necessary link

between an insured peril (also known as insured risk) and

the insured harm (for example, damage, personal injury

or business interruption), and also between the harm

and consequential loss (in policies such as business

interruption policies where such loss is recoverable). The

issue of causation was explored in a number of important

cases arising out of non-damage business interruption

coverage in the context of COVID-19, centering on the

Supreme Court test case, Financial Conduct Authority v

Arch Insurance (UK) Ltd and others [2021] UKSC 1 (FCA

v Arch (SC)), decided in early 2021 (after an expedited

trial in the High Court (FCA v Arch Insurance (UK) Ltd and

others [2020] EWHC 2448 (Comm) (FCA v Arch (HC)) and

leapfrog appeal). The ﬁndings of the Supreme Court in

FCA v Arch (SC) relating to causation, although made in

the context of COVID-19 business interruption insurance

claims, have wider consequences as they are relevant to

all types of insurance claims where there are concurrent

causes of loss.

This note explains:

• The principle of proximate cause in general insurance

contract law.

• The relevance of the “but for” test of causation when

determining proximate cause.

• The approach of the Supreme Court in FCA v Arch

(SC) when determining causation where there are

multiple concurrent causes of loss.

• The relevance of causation when quantifying the

insurance indemnity.

For a summary of the decisions in FCA v Arch (SC) and

FCA v Arch (HC), see Articles, COVID-19: implications

of Supreme Court judgment in FCA BI test case and

COVID-19: implications of judgment in business

interruption test case. For an explanation of the

background to the litigation see Practice note, COVID-19:

FCA business interruption insurance test case.

How causation issues can arise

The risk undertaken by an insurer is typically deﬁned by

certain consequences having been caused by certain

insured perils. In perhaps the simplest case, of a

property damage policy, it will be necessary to show that

property damage was caused by an insured peril, which

may be identiﬁed in an exhaustive list of perils (such as

ﬁre and ﬂood) or may be identiﬁed negatively (an “all

risks” policy that covers all risks save for those that are

excluded). The necessary causation enquiry focusses

on the link between the peril and the damage: was the

property damage sufﬁciently caused by an insured peril,

for example, ﬁre?

In some policies the causation enquiry can arise at

different stages and often more than once, as follows:

• Within the insured peril.

• Between the insured peril and the harm.

• Between the harm and the (consequential) loss.

Within the insured peril

Some insured perils are “composite” perils which require

different things to have happened in a causal sequence

(see FCA v Arch (SC) at paragraph 216). Non-damage

business interruption cover (that is cover that expressly

provides an indemnity for interruption in the absence of

damage to the insured property) might be triggered by

“inability to use the insured premises due to restrictions

imposed by a public authority during the period of

insurance following … an occurrence of any human

infectious or human contagious disease”. (For more

2 Practical Law

Reproduced from Practical Law, with the permission of the publishers. For further information visit uk.practicallaw.thomsonreuters.com

Insurance contract law: causation

information, see below, Multiple causal connections

within a single cover).

Between the insured peril and the harm

The connection between the insured peril (for example,

ﬁre or disease) and the harm (damage to property or

business interruption) is the classic and most common

causational connection required in insurance, and is

similar to the classic causational requirement in tort

(between the acts or omissions of the tortfeasor and the

personal injury or property damage).

In relation to property policies, it is necessary to show

that an insured peril caused the physical destruction or

damage, which is the harm to the policyholder’s interest

in the property that is being insured.

Similarly, for business interruption policies, it is necessary

to show that, for example, the damage or disease caused

the interruption. It is clear that interruption is not part

of the description of the insured peril but rather “a

description of the type of loss or damage covered by the

policy, in the same way as the type of loss or damage

covered by, for example, a buildings insurance policy is

physical destruction or damage”. The word interruption

“describes the nature of the harm to the policyholder’s

interest in the subject matter of the insurance for which

an indemnity is given if it is proximately caused by an

insured peril” (FCA v Arch (SC), paragraph 215).

Therefore, in the example of a composite peril given

above, Within the insured peril, it would be necessary to

show that the peril occurred, namely that the inability

to use the premises arose under the speciﬁed causal

sequence (involving disease followed by public authority

restrictions), but also that the business interruption

was itself caused by the inability to use the premises

resulting from those underlying causes.

This causal link between the peril and the harm is

considered in detail below, Proximate cause and the link

between the insured (or excluded) peril and the harm.

Between the harm and the

(consequential) loss

For some types of policy, the concept of “loss” (at

least as a concept of measuring losses at large) is

not relevant. A property policy typically speciﬁes

that, once the necessary harm (physical damage) has

been established, the indemnity is to be calculated

by reference to the cost of reinstatement (repair or

replacement) of the property. Similarly, a liability policy

provides an indemnity against the insured’s liability to

a third party and costs that fall within the cover, so that

quantiﬁcation of the insured’s loss (and the insurance

indemnity) does not engage causation issues in the

same way as other types of cover, such as those which

provide an indemnity for consequential loss.

Consequential loss, such as lost proﬁts, rent and other

ﬁnancial losses, is typically not recoverable unless

described in the policy and insured as such (Re Wright

and Pole (1834) 1 AD & EL 621).

However, some types of policy expressly cover

consequential loss and require that such loss be caused

by a trigger event. Although quantifying that loss is

merely the “pecuniary measure” of the harm (such

as interruption in a business interruption cover), this

quantiﬁcation includes a causation requirement: the

loss must result from (or similar express wording) the

interruption due to an insured peril (FCA v Arch (SC),

paragraph 216). The quantiﬁcation machinery, which

in business interruption cover includes “trends or

circumstances” clauses in the policies (see below, Trends

clauses), further emphasises that the aim is to identify

losses caused by the interruption, as distinct (in business

interruption policies, for example) from losses caused by a

downturn in revenue which would have occurred because

of business trends or other circumstances independent of

the interruption and the insured peril which gave rise to it.

Other kinds of policy than business interruption may

also involve claims for consequential loss, such as

insurance in respect of rent, hire-purchase transactions

and insurance for the cancellation of events such as

concerts or ﬁlms (see MacGillivray on Insurance Law

(Sweet & Maxwell, 15th edition), Chapter 31, paragraphs,

33-015 and 33-016).

Multiple causal connections within a single

cover

The causation requirement for the purpose of

quantifying loss is illustrated by the “hybrid” clause (a

clause that provides cover for a combination of disease

and prevention of access/public authority action)

included in one of the policies issued by insurer Hiscox

and considered in FCA v Arch (SC) (paragraph 216):

”Setting out the elements of the insured peril

in their correct causal sequence, they are:

(A)an occurrence of a notiﬁable disease, which

causes(B) restrictions imposed by a public

authority, which cause (C) an inability to use

the insured premises, which causes (D) an

interruption to the policyholder’s activities that is

the sole and direct cause of ﬁnancial loss. Counsel

for Hiscox in their submissions on this issue

usefully represented the structure of the clause

in a symbolic form as A→B→C→D, where each

arrow represents a causal connection.”

In that analysis, the insured peril is the “causal sequence”

A→B→C (all of which must occur with those necessary

3 Practical Law

Reproduced from Practical Law, with the permission of the publishers. For further information visit uk.practicallaw.thomsonreuters.com

Insurance contract law: causation

causal connections), the harm that must be caused by

the peril is D (interruption), and further it is necessary for

loss to be (solely and directly in this case) caused by that

interruption.

Proximate cause and the link

between the insured (or excluded)

peril and the harm

Section 55 of the Marine Insurance

Act 1906 and the role of contractual

intention and construction

The core link between the insured peril (or, indeed,

although the point arises less frequently, any excluded

peril) and the relevant harm is the proximate cause test.

This is set out in section 55 of the Marine Insurance Act

1906 (MIA 1906) (although is accepted as applicable to

non-marine insurance) in the following terms:

“… unless the policy otherwise provides, the

insurer is liable for any loss proximately caused

by a peril insured against, but, subject as

aforesaid, he is not liable for any loss which is not

proximately caused by a peril insured against.”

As the opening words of section 55 make clear, the

requirement is subject to contrary provision in the

relevant policy. That is, the insurance policy may provide

for a different connection between the insured (or

excepted) peril and the loss. For example, the words

“caused by” have historically and uniformly been

interpreted as importing the concept of proximate

cause. By contrast, the words “caused directly or

indirectly by” have been interpreted to refer to causes

which are more immediate or more remote than the

proximate cause (Brian Leighton (Garages) Limited v

Allianz Insurance Plc [2023] EWCA Civ 8, at paragraph

29. See Legal update, Consideration of “damage caused

by pollution or contamination” exclusion in insurance

policy (Court of Appeal)).

The term proximate causation is not deﬁned in the MIA

1906, but the case law establishes that the requirement

is based on the presumed intentions of the parties,

and therefore is a question of construction (FCA v

Arch (SC) paragraphs 163 and 190, summarising some

of the authorities). As Lord Briggs (in the minority

in FCA v Arch (SC) but in agreement on this point)

noted at paragraph 320: “The question whether

particular consequential harm to a policyholder is

subject to indemnity is as much a part of the process of

interpreting their bargain as is the identiﬁcation of the

insured peril.” In other words, the question is whether,

on the proper construction of the policy, the parties

intended that the insurance would respond to harm

causally related to the insured peril in the manner that it

was in fact (see below, The proximate cause test).

The proximate cause test

Unless the policy provides otherwise, the insurer is

liable for any loss proximately caused by an insured peril

(section 55, MIA 1906 and FCA v Arch (SC) at paragraph

162). This is sometimes said to be a “fundamental

rule of insurance law” (see for example MacGillivray

on Insurance Law, Chapter 19, paragraph 19-001).

Ultimately, however, the proximate cause test rests

on the presumed intention of the parties (see Leyland

Shipping Co Ltd v Norwich Union Fire Ins Sy Ltd [1918] AC

350, per Lord Atkinson at 365 and Lord Shaw at 369,

quoted in FCA v Arch (SC) at paragraph 166).

The speeches in Leyland Shipping used a number of

descriptions of proximate cause but Lord Shaw provided

the following guidance (pages 369 and 370):

• The true and overriding principle is to look at a contract

as a whole and to ascertain what the parties to it really

meant.

• The proximate cause is not the cause proximate in

time to the loss, but rather “that which is proximate in

efﬁciency.”

• Causation is not a chain but a net and “where

various factors or causes are concurrent, and one

has to be selected, the matter is determined as one

of fact, and the choice falls upon the one to which

may be variously ascribed the qualities of reality,

predominance, efﬁciency.”

• At any given point in a net, inﬂuences, forces and

events converge from all directions, not just in a

straight line and an earlier cause may be more potent

that a later one (Allianz v University of Exeter [2023]

EWHC 630 (TCC)).

Although the authorities often refer to “the” proximate

cause or “the” real efﬁcient cause of loss, in reality, a

loss may result from a combination of causes, either

operating independently of one another, or, often, in

a chain where each would not have arisen but for that

preceding it in the chain. Of these causes, the search

is for the, or a, proximate cause and it is generally

irrelevant if a cause is either more remote in the chain

than the proximate cause, or more immediate (Reischer

v Borwick 1894 2 QB 548, Brian Leighton (Garages)

Limited v Allianz Insurance Plc [2023] EWCA Civ 8 and

Allianz v University of Exeter [2023] EWHC 630 (TCC),

where the court held that the insurer was entitled to

reject a claim on the basis that the proximate cause of

the damage was the dropping of a bomb during the

Second World War (a peril that was excluded under

a war exclusion in the policy) and not the controlled

detonation of the bomb decades later in 2021. See

4 Practical Law

Reproduced from Practical Law, with the permission of the publishers. For further information visit uk.practicallaw.thomsonreuters.com

Insurance contract law: causation

also, below The effect of the passage of time on

proximate causation).

Determining the real or dominant cause of loss has been

said to require the application of business or relevant

industry common sense (see Leyland Shipping per Lord

Dunedin at page 362, The TM Noten BV v Harding [1990]

2 Lloyd’s Rep 283 per Bingham LJ at pages 286 and

287 and Global Process Systems Inc v Syarikat Takaful

Malaysia Bhd (The Cendor Mopu) per Lord Savile at

paragraph 19). However, in FCA v Arch (SC) the Supreme

Court sounded a note of caution, explaining that the

common sense principles or standards to be applied in

selecting an efﬁcient cause of the loss “are capable of

some analysis,” and that it is “not a matter of choosing

a cause as proximate on the basis of an unguided gut

feeling” (paragraph 168).

Policy language

The meaning of proximate causation is not ﬁxed and

ultimately depends on the speciﬁc construction of the

particular contract. Typically, there will be a term used

in the policy to describe the requisite link between the

peril and the harm. That term may be “caused by”, but

it may be something else. The policy may require that

loss or damage be “suffered by”, for example, ﬁre or

hurricane. The interruption may need to “result from”

or be “in consequence of” public authority action. Other

commonly used terms are “following” or “arising out of”,

and sometimes with qualiﬁcations such as “directly” or

“directly or indirectly”, or “solely”.

As explained above, The proximate cause test, the

default position is that the applicable test is proximate

causation unless the contract provides with sufﬁcient

clarity that something else is to apply. Although, there

are bodies of law on the meaning of speciﬁc words,

the Supreme Court in FCA v Arch (SC) noted that while

words such as “indirectly” or possibly “following” may

indicate a looser causal requirement than proximate

cause, “it is rare for the test of causation to turn on such

nuances”. The Supreme Court’s conclusions in that

case (that on the proper interpretation of the disease

clauses, in order to show that business interruption loss

was proximately caused by one or more occurrences of

illness resulting from COVID-19, it was sufﬁcient to prove

that the interruption was a result of government action

taken in response to cases of disease which included at

least one case of COVID-19 within the geographical area

covered by the clause) applied across a range of policy

wordings, including “following”, “arising from” and “as a

result of” (paragraphs 162 and 212 of the judgment).

As to the word “indirectly” broadening the proximate

cause test (indicating that the causative link may be

more remote than a proximate cause), in Crowden v QBE

Insurance (Europe) Ltd [2017] EWHC 2597 (Comm) Peter

McDonald Eggars QC stated that even an indirect cause

“must be signiﬁcant; it must stand out as a contributing

factor, at least, to the claim, liability or loss” (see

paragraphs 71 and 72, where the judge cited Scrutton

J’s words in Coxe v Employers’ Liability Assurance

Corporation [1916] 2 KB 629 that he was “unable to

understand what is an indirect proximate cause” and

in his judgment the only possible effect which could be

given to those words was that “a more remote link in the

chain of causation is contemplated than the proximate

and immediate cause”). The phrase “in connection

with” can denote a looser (”relatively weak”) causal

connection (Stonegate Pub Company Ltd v MS Amlin

Corporate Member Ltd and others [2022] EWHC 2548

(Comm) per Butcher J at paragraph 116, considered in

Legal update, Causation, aggregation and furlough

payments considered in insurance claim for business

interruption losses (High Court)).

As to the word “following”, this has been held to import

more than merely a temporal relationship but not

necessarily one of proximate causation (see FCA v Arch

(HC) at paragraph 95). The word “resulting” has been

held to be more consistent with a proximate cause

requirement than the use of the “less forceful” word

“following” (see the Irish decision of Hyper Trust v FBD

Insurance plc [2021] IEHC 178 at paragraphs 174 and

175). The words “occasioned by” have been treated by

agreement of the parties as imposing the proximate

cause standard (Allianz v University of Exeter [2023]

EWHC 630 (TCC), paragraph 18).

The burden of proof

The insured has the burden of proving to the ordinary

civil standard that a loss occurred which was proximately

caused by an insured peril (Rhesa Shipping v Edmunds

(The Popi M) [1985] 1 WLR 948).

However, where the insured has proved a loss resulting

from an insured peril, the insurer then has the burden of

proving the applicability of any exception, condition or

warranty (Colinvaux and Merkin’s Insurance Contract Law

(Sweet & Maxwell, Chapter B12, paragraph B-1080 and

Munro, Brice & Co v War Risks Assoc Ltd [1918] 2 KB 78,

pages 88 and 89 (although the decision was reversed

on appeal, the ﬁndings in relation to the burden of proof

were not considered by the Court of Appeal).

Under an all risks policy, the insured satisﬁes the

applicable burden by proving a loss caused by a

fortuitous event, and the insurer then has the burden

of proving that the loss was caused by an excluded or

uninsured peril (British Marine v Gaunt [1921] 2 AC 41,

per Lord Sumner at 57 and 58).

Generally, the burden of proving that a loss was not

accidental (for example, that it was caused by fraud, or

5 Practical Law

Reproduced from Practical Law, with the permission of the publishers. For further information visit uk.practicallaw.thomsonreuters.com

Insurance contract law: causation

arson) lies on the insurer (Slattery v Mance [1962] 1 QB

676, per Salmon J at 681). However, where an insured

seeks to recover for a peril of the sea (the deﬁnition

of which includes an element of fortuity) it has the

burden of proving that the loss was not caused by its

own deliberate and wrongful actions (The Alexion Hope

[1988] 1 Lloyd’s Rep 311; Brownsvill Holdings v Adamjee

Insurance, The Milasan [2000] 2 Lloyd’s Rep 458 and

Suez Fortune Investments Ltd v Talbot Underwriting Ltd

(Brillante Virtuoso) [2019] EWHC 2599, per Teare J at

paragraph 60).

The burden of proof may be displaced by express terms

of the policy (Levy v Assicurazione Generali [1940] AC

791, where the policy expressly provided that where

the insurer sought to rely on a policy term such as an

exclusion to reject the claim, the burden of proving

that the loss was covered was on the insured. However,

Mustill J adopted a more cautious approach to terms

reversing the burden of proof in Spinney’s (1948) Ltd v

Royal Insurance Co [1980] 1 Lloyd’s Rep 406, where it was

held that insurers still had to produce evidence from

which it could reasonably be argued that the relevant

exclusion applied).

The effect of the passage of time on

proximate causation

The proximate cause test described above, The proximate

cause test, is more sophisticated than an investigation

merely into what was the most recent cause. In some

cases, the proximate cause may be the most recent cause,

but in others it will not be. The fact that timing does not

determine what cause is proximate is well illustrated by

JJ Lloyd Instruments Ltd v Northern Star Insurance Co Ltd

[1987] 1 Lloyd’s Rep 32 (”The Miss Jay Jay”) in which the

Court of Appeal held there were two proximate causes

of the loss of a vessel, ﬁrst, its inherent unseaworthiness

(due to design defects), and second the (obviously

subsequent) adverse sea conditions experienced on a

voyage. Further, in Allianz v University of Exeter [2023]

EWHC 630 (TCC), the proximate cause (act of war) arose

many decades prior to the most recent cause (controlled

detonation, some 80 years later).

A particularly stark issue in relation to the passage

of time arose in the COVID-19 business interruption

cases. The FCA v Arch test case concerned only cases of

COVID-19 from the start of 2020 until the ﬁrst lockdown

and related public authority orders in late March 2020.

There was no explicit consideration in the test case as

to whether any COVID-19 cases were too early in 2020

to qualify as proximate causes of the March 2020

government actions. The Supreme Court simply held,

in agreement with the analysis of the Divisional Court,

that the proximate cause test was satisﬁed if there had

been at least one occurrence of COVID-19 within the

geographical area covered by the relevant clause which

predated the government action, explaining that “each

of the individual cases of illness resulting from COVID-19

which had occurred by the date of any Government

action was a separate and equally effective cause of

that action (and of the response of the public to it)”

(FCA v Arch (SC), paragraph 212). There was also no

consideration whether any government action after the

ﬁrst lockdown was lifted in July 2020 was proximately

caused by occurrences of COVID-19 prior to the ﬁrst

lockdown, or only proximately caused by later cases

more recent to the government actions in the second

half of 2020 (and beyond). Both questions arose in

subsequent cases.

The ﬁrst question was due to be considered in the

London International Exhibition Centre and other

test cases in spring 2023 but it was not pressed as a

preliminary issue and so, although pleaded, remains to

be determined in a later trial.

The second question was considered in a collection

of test cases on aggregation and causation heard

consecutively in 2022 before judgments were given in

any, namely Stonegate Pub Company Ltd v MS Amlin

Corporate Member Ltd and others [2022] EWHC 2548

(Comm), Various Eateries Trading Ltd v Allianz Insurance

plc [2022] EWHC 2549 (Comm) and Greggs plc v Zurich

Insurance plc [2022] EWHC 2545 (Comm).

In Stonegate and Various Eateries, the policyholders

argued that there was cover for business interruption

loss occurring after the period of insurance had ended

but within the maximum indemnity period (MIP)

speciﬁed in the policy. In Stonegate, for example,

they argued that their losses in the period 30 April

2020 (when the period of insurance ended) to 30

April 2023 (the end of the MIP) were proximately

caused by covered events (namely cases of COVID-19

occurring within the Vicinity (as deﬁned) and within

the period of insurance). They contended that, based

on the ﬁndings relating to causation in FCA v Arch

(SC), the occurrences of COVID-19 within the period

of insurance were concurrent proximate causes of the

cases of COVID-19 that occurred after the insurance

had expired. The court held that FCA v Arch (SC) had

not decided that each case of COVID-19 was equally

causative of government or consumer response over a

prolonged period of time. It was not plausible that early

cases of COVID-19 (that is those occurring before April

2020) were equal or approximately equal causes of

the various government measures adopted at different

stages during the MIP and of consumer behaviours at

the different times during that period. The incidence

of cases varied over time and with that variation

there were changes in governmental and consumer

responses (paragraphs 201 to 204 of the judgment).

6 Practical Law

Reproduced from Practical Law, with the permission of the publishers. For further information visit uk.practicallaw.thomsonreuters.com

Insurance contract law: causation

The reasoning in Stonegate was also applied in Various

Eateries, where the judge further stated that the fact

that the cases of COVID-19 occurring in the period of

insurance may have caused later cases of the disease

(because cases make cases) was not sufﬁcient to say

that the cases of the disease in the period of insurance

were the proximate cause of the governmental

measures and public response after the period of

insurance. However, insurers accepted that cases of

COVID-19 before the period of insurance ended could

have caused interruption after the period of insurance

ended. Examples could be cases where people had

died as a result of contracting COVID-19 before the end

of the period of insurance, or cases of long COVID in

people who were infected before the end of the period

of insurance, but it was for the policyholders to prove

that there had been interruption as a result of COVID-19

cases before the period of insurance ended (paragraphs

61 and 62 of judgment). These causational ﬁndings in

Stonegate and the related cases will not form part of the

appeal to the Court of Appeal that is due to take place.

Multiple proximate causes

Where there are two (or more) proximate causes of a

harm, one covered, but the (or an) other excluded, the

exclusion prevents recovery (Wayne Tank & Pump Co

Ltdv Employers Liability Assurance Corporation Ltd [1974]

QB 57 (per Cairns LJ); The Demetra K [2002] 2 Lloyd’s

Rep 581 (per Lord Phillips at paragraph 18); Global

Process Systems Inc v Syarikat Takaful Malaysia Bhd (The

Cendor Mopu) (per Lord Savile at paragraph 22 and Lord

Mance at paragraph 88); Navigators Insurance Co Ltd v

Atlasnavios-Navegacao Lda, [2018] UKSC 26 (per Lord

Mance at paragraph 49, considered in Legal update,

Supreme Court upholds decision that war risks insurers

not liable to shipowners); and FCA v Arch (SC) at

paragraphs 172 and173). In FCA v Arch (SC), it was said

that, although it is “always a question of interpretation”,

where there are two proximate causes, one of which is

an insured peril but the other is excluded “the exclusion

will generally prevail” (paragraph 174).

Conversely, if one proximate cause is covered, and the

other is neither covered nor excluded, the policy will

respond (The Miss Jay Jay, especially per Slade LJ at

page 40).

In FCA v Arch (SC), the key hurdle for the policyholders

was the “but for” test, that is showing, for example,

that “but for” the cases of COVID-19 in the vicinity they

would not have suffered any interruption (see below,

Satisfying the “but for” test is not enough to satisfy

proximate causation), but the reasoning of the Supreme

Court depended upon the ﬁnding that the insured peril

was one of multiple broadly equal effective/dominant

causes. This applied to the case of disease within the

radius as regards the disease clauses (paragraphs 189

to 191, 212, 295, 319, 321). This was possible because,

although there were very many (perhaps a million, said

Lord Briggs at paragraphs 319 and 321) concurrent

causes, none was more dominant than another

(paragraph 212).

It similarly applied, with hybrid (which combine disease

and prevention of access) and prevention of access

clauses as regards the insured public authority ordered

closure, and the concurrent other causes of interruption

such as the advice that customers stay at home, fear or

illness (paragraphs 229 and 230 and 237 to 239).

Where the insured peril was not even an equally

dominant cause (for example, where the prevention of

access really had little to do with the interruption) then

there would be no cover, even despite the Supreme

Court conclusions relating to the “but for” test

(paragraph 244, see also below, Satisfying the “but for”

test is not enough to satisfy proximate causation).

In particular, what is unlikely to be required or permissible,

on the proper construction of a particular policy, in cases

of multiple concurrent proximate causes is to weigh

the totality of insured perils against the totality of the

uninsured perils, providing all the perils are individually

sufﬁciently effective to be proximate causes (FCA v Arch

(SC) paragraphs 198 and following). This is not the same,

of course, as dividing up causes where the harm or loss

can itself be apportioned with some exclusively caused

by one or more causes and other harm or loss caused by

other causes (FCA v Arch (SC) paragraph 198). If such

apportionment is possible in a particular case, the losses

due to insured causes would be covered, whereas those

due to uninsured causes would not (Stanley v Western

Insurance Company (1868) LR 3 Ex 71, 74-75).

Summary

Where there are more than one proximate

causes, if one is an excepted peril then there will

be no cover, but otherwise there will be cover

providing at least one proximate cause is an

insured peril (subject to discussion of the but

for test immediately below) even though other

proximate causes are not insured perils.

Satisfying the “but for” test is not

enough to satisfy proximate causation

The basic justiﬁcation for the proximate cause test is

that it is not sufﬁcient to show that a harm would not

have occurred “but for” the insured peril (MarsdenvCity

and County Assurance Co (1865) LR ICP 232). For

7 Practical Law

Reproduced from Practical Law, with the permission of the publishers. For further information visit uk.practicallaw.thomsonreuters.com

Insurance contract law: causation

insurance cover purposes, the “but for” test alone

returns a lot of “false positives” (FCA v Arch (SC)

paragraph 181). Insurance policies (like legal causation

in contract and tort) typically require more, and that is

what proximate cause does, by asking whether of all the

causes in the world that together contribute to bringing

about the harm, the insured peril is a sufﬁciently

effective cause for the policy to respond.

Is it necessary to satisfy the “but

for” test in order to satisfy proximate

causation in multiple concurrent cause

cases?

There is a separate question as to whether it is necessary

(even though not sufﬁcient) to satisfy the “but for” test.

Usually, it is.

When considering the link between a wrongdoing and

loss in tort or contract law, it is fundamental that the

loss would not have occurred “but for” the wrongdoing.

This is part of the very measure of loss in tort and

contract law, which requires the claimant to be put in

the position it would have been put in had the tort or

contract breach not occurred. And, as set out below,

Trends clauses, trends clauses in business interruption

policies (being clauses that allow for adjustments to

proﬁt ﬁgures to be made to take account of any trends

and other circumstances which would have affected

the business even in the absence of the insured event)

similarly make explicit the “but for” requirement of

causation of loss. This is therefore core to the basic

concept of what it is to cause something in most

situations.

When considering the link between the insured peril

and the harm in an insurance policy, there is no such

explicit requirement in section 55 of the MIA 1906

(which requires that loss is “proximately caused”).

Insurance, unlike contract and tort, is not about blaming

the insurer for what happened, but merely about

working out whether the risks taken by the insurer (the

insured perils) are sufﬁciently linked to the harm for

cover to respond. But in that context, the “but for” test

will usually need to be satisﬁed. As the Supreme Court

pithily explained in FCA v Arch (SC) at paragraph 181:

”We agree with counsel for the insurers that in

the vast majority of insurance cases, indeed in

the vast majority of cases in any ﬁeld of law or

ordinary life, if event Y would still have occurred

anyway irrespective of the occurrence of a prior

event X, then X cannot be said to have caused Y.”

Usually, if a harm (such as physical damage to property

or interruption to a business) would have occurred

even “but for” the insured peril, the harm will not fall

within the insured risk. The insurer can usually say that

in covering the insured against damage resulting from

ﬂoods, it was not intended to cover damage that would

have occurred even without a ﬂood.

However, as was observed in FCA v Arch (SC), referring to

some of the tort case law and academic commentary and

examples, in some cases it is contrary to common sense

to apply a “but for” test, especially where there is a set of

concurrent causes none of which individually were “but

for” causes of the relevant consequence (that is, without

any one cause the consequence would have occurred

unchanged) but where the set together was necessary

(that is, without the set of causes the consequence

would not have occurred). In those circumstances, the

contribution that any of these causes makes - a lesser

causal connection than “but for” - is enough causal

connection (paragraphs 182 to 185). Examples of where

the “but for” test is inadequate to determine causation,

include where two hunters both fatally shoot a victim

(a case where both were sufﬁcient - that is, each alone

would have led to the result, yet because of the presence

of the other neither was necessary as the harm would

have happened even “but for” each when considered

alone (an example derived from Cook v Lewis [1952] 1

DLR 1(a tort case)); or 20 people combine to push a bus

over a cliff (a case where no individual was a necessary or

sufﬁcient cause).

It is important here to distinguish the problem considered

in the Fairchild v Glenhaven Funeral Services Ltd [2002]

UKHL 22 line of cases relating to injury caused by the

negligent exposure to asbestos. There, on the evidence,

only one of the multiple contender causes was a “but for”

cause of the loss (a single asbestos ﬁbre) but the state of

science meant that there could never be enough evidence

to prove which one, that is, it is a problem of evidence and

proof and not a problem with the “but for” test.

One of the main contributions of the decision in FCA v

Arch (SC) is to make clear that the question of whether

the proximate cause requirement, or, more accurately,

the particular causal connection required by the express

words of a particular policy, is satisﬁed in a particular

case of concurrent causes where none are “but for”

causes, is a question of construction of the policy

(paragraphs 190 and 191).

An example relied on in FCA v Arch (SC) relates to

insurance decisions concerning defence costs, such as

International Energy Group Ltd v Zurich Insurance plc

UK Branch (Association of British Insurers and another

intervening) [2015] UKSC 33 and the Privy Council

decision in New Zealand Forest Products Ltd v New

Zealand Insurance Co Ltd [1997] 1 WLR 1237, where it

was held or accepted that insureds are entitled to an

indemnity even though the “but for test” is not satisﬁed

(see the majority at paragraph 186 to188 although Lord

Briggs, speaking for the minority, warned they may be

8 Practical Law

Reproduced from Practical Law, with the permission of the publishers. For further information visit uk.practicallaw.thomsonreuters.com

Insurance contract law: causation

sui generis (paragraph 326)). In those cases, the policy

covered defence costs attributable to covered litigation

against the insured even if those costs would have been

incurred without that litigation because they were jointly

incurred for the purposes of other litigation that was not

insured (that is, they were necessary for the purpose of

dealing with, for example, uninsured claims against the

insured arising out of fraud allegations, or uninsured

claims against uninsured co-defendants). This is simply

because the case law conﬁrms that such policies are

intended to cover those costs (and that such costs do

“arise from”, “on account of”, or similar, the covered

litigation) even though the insured peril is not a “but

for” cause of the costs.

The Supreme Court’s approach

to the “but for” issue on the

facts of FCA v Arch

Disease clauses

The primary concurrent causation issue relating

to the disease clauses in the FCA v Arch test case

was that the insured peril for such clauses was

conﬁned to the occurrence of disease within 25

miles or 1 mile of the insured premises, but the

interruption resulted from national government

measures which were a response to the disease

nationwide and would have happened even

without the cases in any particular 1 mile or 25

mile radius (see for example paragraph 179).

Accordingly, even if any speciﬁc occurrence,

or even if all cases within a circle around the

premises with a 1 mile or 25 mile radius, had

not taken place, the national measures and

so interruption would still have occurred. The

insured peril was therefore not a “but for”

cause of the interruption, and the question of

construction arose as to whether it therefore was

or was not a cause of the interruption within the

meaning of the causal requirements speciﬁed in

the policy (see paragraph 192).

Insurers maintained that the cover in practice

only or mostly responded to local outbreaks that

were entirely or mostly within the relevant radius,

such that any public authority action was “but

for” (and dominantly) caused by the occurrences

within the radius. Accordingly, the “but for”

test prevented cover in the circumstances of a

national government response to a nationwide

pandemic, as that was in partly responding to the

local cases.

The Supreme Court construed the relevant

policies emphasising that many of the diseases

to which the insured peril related (notiﬁable

diseases, including SARS) were of their nature

infectious diseases that can spread widely, and

so the parties must have contemplated that

the disease could occur inside and outside the

radius, and that the entire outbreak could trigger

the relevant public authority action (paragraphs

194 to 197 and the minority at paragraphs 315

and 316). Further, the entire outbreak did in fact

do so, and each case of COVID-19 was an “equal

and effective cause” of the government action

and public response to it (paragraph 212. See

also the Irish decision of Hyper Trust Ltd v FBD

Insurance plc [2021] IEHC 78 at paragraphs 190

and 198 to 199). In those circumstances, “the

parties could not reasonably be supposed to

have intended that cases of the disease outside

the radius could be set up as a countervailing

cause which displaces the causal impact of the

disease inside the radius” at least where the

insurers do not expressly conﬁne the insured

peril to interruption solely or only caused by (/

resulting from/following, etc.) the disease within

the radius, and had not speciﬁed an exclusion

to that effect (FCA v Arch (SC), paragraph

195). The causal requirements therefore do

not require that the “but for” test be satisﬁed

in these circumstances in which the cases

within the radius contributed to the national

action (paragraph 212). (See further the similar

analysis of the Irish High Court in Hyper Trust at

paragraphs 143 to 147.)

Therefore, it was held that the logic of the

examples of the two hunters or (more closely)

20 bus-pushers be applied to concurrent

effective but not “but for” causes, even if (as in

this case) there were thousands or even over

a million such concurrent causes (namely, the

individual occurrences of COVID-19 to which

the government’s nationwide instructions and

legislation were a reaction) (paragraphs 189

to191 and 319).

It is important to understand that the

requirement for proximate causation is not

being disapplied here. There is still a causal

requirement, however that requirement is

satisﬁed without the “but for” test to the

extent of other concurrent COVID-19-related

causes that also do not satisfy the “but for”

test. Concurrent causes unrelated to COVID-19,

9 Practical Law

Reproduced from Practical Law, with the permission of the publishers. For further information visit uk.practicallaw.thomsonreuters.com

Insurance contract law: causation

however, such as a chef who was due to

leave anyway, will prevent or reduce cover

(paragraphs 231 and 232 (or if a licence would

not have been renewed in any case (Hyper Trust

at paragraphs 205 and 221).

Prevention of access and hybrid

clauses

Similar issues arise for prevention of access and

hybrid clauses (as those terms were used in the

FCA v Arch test case). These involved composite

perils such as those referred to above, Within

the insured peril. They raise further “but for”

problems. Where the insured peril includes a

requirement of prevention of access/hindrance

of use/inability to use/closure of the premises

(or similar), there is a question whether the,

or at least some, interruption (and, later, the

loss) was caused by the prevention of access

to the premises, in circumstances in which,

even without that prevention of access, there

would have been interruption or loss because

of the broader effects of COVID-19 beyond that

prevention of access to the premises. Those wider

effects of COVID-19 include the stay at home

and distancing instructions (on customers and

employees), self-imposed fear and illness, the

closure of surrounding businesses (which itself

is sometimes covered by “loss of attraction”

business interruption extensions), and the general

business downturn during the pandemic. Indeed,

such concurrent causation is shown by the fact

that the businesses suffered a substantial drop in

income as a result of COVID-19 even before they

were ordered to close (that is, before the insured

peril of prevention of access was triggered).

The Supreme Court was clear that, for example,

where the government orders a shop to close, all

COVID-19 losses (from the walk-in business) are

recoverable, not only those losses which would

not have occurred had the shops remained

open but COVID-19 still existed. The closure (an

insured peril) and stay at home orders (not an

insured peril) are concurrent “but for” causes

and so either satisﬁes the necessary causal

requirements. The consequences of COVID-19,

which would have caused business interruption

loss even in the absence of the insured peril

(prevention of access), must be removed from

the counterfactual when considering whether

the interruption was caused by the prevention of

access and what loss is recoverable (paragraphs

229 to 230 and 247).

This would also be true for other prevention of

access-type clauses, such as vermin clauses: if

a policyholder takes out cover for interruption

due to an authority action following discovery

of vermin, it would not be reasonably

understood (without express words to the

contrary) that the indemnity would be limited

to loss that would not have been suffered

but for the forced closure of the premises

(that is it would not be open to argue that a

vermin infestation would have led to business

interruption loss even in the absence of public

authority action) (paragraphs 238 to 239). The

Supreme Court also drew an analogy with the

facts in IF P&C Insurance Ltd v Silversea Cruises

[2004] EWCA Civ 769, where the cover was for

business interruption loss resulting from US

State Department warnings following the 9/11

attacks, and the indemnity was not reduced to

the extent that interruption and loss would still

have resulted from the attacks even without the

State Department Warnings (FCA v Arch (SC)

paragraphs 241 and 242).

This solution to the “but for” problem,

namely that on their proper interpretation,

the prevention of access clauses covered loss

caused by the orders to close regardless of

whether there were other concurrent uninsured

(but not excluded) causes of the loss, is not

the end of the enquiry. There will still be cases

in which the prevention of access was not a

sufﬁciently dominant cause, as compared with

the other consequences of COVID-19, for it to be

a proximate cause at all. This was explained in

FCA v Arch (SC) (paragraph 244):

“For completeness, we would point

out that this interpretation depends

on a ﬁnding of concurrent causation

involving causes of approximately equal

efﬁcacy. If it was found that, although

all the elements of the insured peril

were present, it could not be regarded

as a proximate cause of loss and the

sole proximate cause of the loss was

the COVID-19 pandemic, then there

would be no indemnity. An example

might be a travel agency which lost

almost all its business because of the

travel restrictions imposed as a result

of the pandemic. Although customer

access to its premises might have

become impossible, if it was found that

the sole proximate cause of the loss of

10 Practical Law

Reproduced from Practical Law, with the permission of the publishers. For further information visit uk.practicallaw.thomsonreuters.com

Insurance contract law: causation

its walk-in customer business was the

travel restrictions and not the inability of

customers to enter the agency, then the

loss would not be covered.”

The same will apply once the prevention ceases

over time to be an effective cause of loss because

the prevention has ceased, the premises are

open, and the consequences of the closure

have abated. One would not continue to strip

out the consequences of COVID-19 then, when

prevention of access has ceased to have any

effect (Hyper Trust at paras 215, 262 to 266).

With “hybrid” clauses, which are a hybrid of

disease and prevention of access clauses, the

composite insured peril requires prevention

of access or similar to have been caused by

occurrences of a notiﬁable disease within

25 miles or similar. The issue in relation to

disease clauses, as to whether the disease

within 25 miles caused the national action

which caused the order preventing access to

the premises, arises, although here it does so

not as to the link between the insured peril

and the harm of interruption, but as to a link

within the composite peril between elements

of it (prevention of access and occurrences of

disease). The issue in relation to prevention of

access clauses also arises, as to whether the

prevention caused interruption or loss that

would have happened even without the order

preventing access due to other COVID-19 effects.

Unsurprisingly, these issues were resolved by the

Supreme Court in the same way as for disease

and prevention clauses.

The “underlying fortuity”/”

originating cause” basis

The Supreme Court explained its conclusions

on to the application of the “but for” test and

the principle of concurrent cause, on the basis

that other effects of COVID-19, “although

not themselves covered by the insurance, …

are matters arising from the same original

fortuity which the parties to the insurance

would naturally expect to occur concurrently

with the insured peril. They are not in that

sense a separate and distinct risk” (paragraph

237). The “originating cause” of the insured

peril - the cause that comprises or sits behind

the first element of the insured peril - was the

global COVID-19 pandemic, and the parties

did not intend any consequences of that

originating cause to restrict the scope of the

indemnity (paragraphs 240, 247, 284, 294,

295, 309 and 310). (See also Hyper Trust at

paragraph 212, where the Irish High Court

referred to COVID-19 as the “common thread”

between the insured peril and society’s

reaction to the virus.)

This language of “originating cause” is common

in aggregation clauses, where it is generally

used by insurers (or, as applicable, reinsurers)

to gather together multiple incidents or claims

under a single cover limit (see Axa Reinsurance

(UK) Ltd v Field [1996] 1 WLR 1026, and Practice

note, Reinsurance: an overview). The adoption

of the term “originating cause” by the Supreme

Court in considering causation in insurance law,

cannot be accidental. By doing so, the Supreme

Court makes use of the fact that insurers

think of risks in terms of “originating cause”

packages (shown by their sometimes providing

aggregation clauses for them), and in future

they can now also sometimes expect to cover

all the consequences of such causes where:

• Those consequences are concurrent causes of

harm;

• None are “but for” causes of the harm;

• At least one of them is an insured peril; and

• None are excluded.

It is also worth noting some ambiguity as

to the geographic extent of the underlying

fortuity (COVID-19) identified in FCA v Arch

(SC). The UK national action was only a

response (or only primarily a response) to

COVID-19 within the UK, but that is primarily

an issue relevant to the question of whether

the interruption or prevention was caused

by the disease within 25 miles of the insured

premises. As to the broader question of

what COVID-19, UK COVID-19 or global

COVID-19, and its consequences should be

stripped out of the counterfactual (that is

what concurrent effects of COVID-19 on the

insured business should be ignored when

determining business interruption loss) the

Supreme Court judgment and declarations

suggest that it is the concurrent effects of the

global COVID-19 pandemic that should be

stripped out (paragraph 240 of the judgment

and declaration 11.1). (See also Hyper Trust at

paragraph 224 as to this uncertainty).

11 Practical Law

Reproduced from Practical Law, with the permission of the publishers. For further information visit uk.practicallaw.thomsonreuters.com

Insurance contract law: causation

Concurrent causes and business

interruption caused by property

damage: the Orient-Express

Prior to FCA v Arch (SC), the leading insurance

case on the “but for” test was the property damage

business interruption case of Orient-Express Hotels Ltd

v Assicurazioni Generali SpA [2010] EWHC 1186. The

policy included a business interruption extension giving

cover for interruption directly arising from damage to

property on an all risks basis. A New Orleans hotel was

damaged by hurricanes Katrina and Rita and, in addition

to the property damage claim, a claim was made on the

business interruption extension. The arbitral tribunal

and then the High Court rejected the claim on the basis

that the insured peril was property damage to the hotel

and even “but for” that property damage there would

have been the same business interruption because the

wider New Orleans area was devastated and evacuated.

In other words, the correct approach was to imagine

a New Orleans devastated by the hurricanes, but with

an unscathed insured hotel in the middle of it, and

to ask what interruption would have been suffered.

(There were also some US cases where the insureds,

rather than the insurers, had argued for this approach,

because in the different circumstances of those cases,

that counterfactual would lead to windfall proﬁts as

the unscathed insured property would have enjoyed

monopoly proﬁts as the only operating business in the

region. In those cases, the approach was generally

rejected. (See FCA v Arch (SC) paragraphs 279 and 280.))

The Divisional Court in FCA v Arch identiﬁed part of the

fallacy in the reasoning of the arbitrators and the judge

as being that:

”they proceeded on the basis that only the

Damage in the abstract should be stripped out

in assessing the counterfactual under the trends

clause [whereas] on a proper analysis, the insured

peril in that all risks policy was not Damage in the

abstract, but Damage caused by a fortuity, there

the hurricane, so that what should have been

stripped out in the counterfactual was not just the

Damage but the Damage and the hurricane” (see

FCA v Arch (DC), at paragraph 345).

The Supreme Court agreed that Orient-Express had

been wrongly decided (despite one of the Supreme

Court panel, Lord Hamblen, having decided that case

in the High Court, and another, Lord Leggatt, having

been on the arbitration panel that decided the same

way), on the basis that both the insured cause (damage

to the hotel) and uninsured causes (damage to the

surrounding area) arose from the same underlying

fortuity (the hurricanes), and although neither satisﬁed

the “but for” test (because of the other), loss resulting

from both causes operating concurrently was covered,

provided (as was the case) that loss from the uninsured

peril was not excluded by the policy (FCA v Arch (SC),

paragraphs 308 to 310).

The Supreme Court did not conduct any more detailed

analysis of the extent to which the parties would have

contemplated that property risks may well affect

a wider area (as it did for notiﬁable diseases), but

presumably such reasoning remains necessary and

implicit. Certainly, it is foreseeable (at least) in the

case of storm damage (an event typically covered by

an all risks policy) that a storm capable of causing

damage to one property will or may cause widespread

damage to other properties. If wider consequences

of the relevant underlying cause were not reasonably

foreseeable it is possible this would lead to a different

result. In other words, there is probably no rule of law or

practice that this result will always apply for concurrent

consequences of an underlying fortuity only one of

which is insured and none of which are “but for” causes:

rather it may always, ultimately, turn on a matter of

contractual interpretation and foreseeability.

Construction of the insured peril

after the FCA v Arch test case

The Divisional Court in FCA v Arch found that the

insured peril was simply the underlying fortuity of

COVID-19, but with a condition requiring that there

had been an occurrence of COVID-19 within the radius

speciﬁed in the policy (for example 25 miles) (paragraph

102 of the judgment and paragraph 64 of the judgment

in FCA v Arch (SC), commenting on this aspect of

the Divisional Court’s decision). The majority of the

Supreme Court rejected this approach, conﬁning the

insured peril to an occurrence of disease (COVID-19)

within the relevant radius; any case of COVID-19

outside that area was not an insured peril (paragraph

74). However, the Supreme Court then held that, as a

matter of construction, the parties intended the causal

requirements to allow recovery for the consequences

of the broader underlying fortuity of COVID-19 (that

is, cover was not conﬁned to business interruption

which resulted only from cases of COVID-19 within the

relevant radius, as opposed to other cases elsewhere).

The minority (Lord Briggs and Lord Hodge), as well

as agreeing with the majority’s approach, would also

have accepted the Divisional Court’s approach on the

basis that it amounted to the same thing in substance.

As Lord Briggs observed, the result of the majority’s

reasoning is that the parties would answer the question

“do clauses with the radius limitations provide cover

for the adverse business consequences of a national

reaction to a national pandemic disease?” in the

afﬁrmative. In their view, this in effect meant that the

12 Practical Law

Reproduced from Practical Law, with the permission of the publishers. For further information visit uk.practicallaw.thomsonreuters.com

Insurance contract law: causation

national pandemic was the insured peril, provided it

reached, spread, encroached or extended within the

radius (paragraphs 322 and 324).

Other radiuses

The FCA v Arch test case only considered 25 mile

and one mile radius disease and prevention of access

clauses. The reasoning focused on the former, and was

then applied to the latter (paragraph 94). (A one mile

prevention of access clause was also considered in Corbin

& King Limited and others v Axa Insurance UK Plc [2022]

EWHC 409 (Comm), considered in Legal update, High

Court holds that COVID-19 related BI losses are covered

by denial of access clause in insurance policy (see below,

Prevention of access clauses for more information)).

Market policies include other distances, including 250

metres and “vicinity of the premises”. The latter was found

to give rise to the same causation test in a prevention of

access wording in Policyholders v China Taiping Insurance

(UK) Co Ltd (10 September 2021) arbitration, considered

in Legal update, COVID-19: arbitrator dismisses business

interruption insurance claims (and see below, Prevention

of access clauses for more information).

”At the premises” disease and similar

business interruption clauses

The FCA v Arch test case did not consider, for example,

clauses requiring the disease or emergency to be “at

the premises” rather than within 25 miles or one mile

of the premises. It was decided in a further set of test

cases (of preliminary issues) under the lead case, London

International Exhibition Centre plc v RSA and others [2023]

EWHC 1481 (Comm) that “at the premises” clauses simply

deﬁned the geographical area in which the occurrence

of the disease had to occur more narrowly than radius

clauses (although some premises are very large), and the

Supreme Court’s causation approach applies equally. In

other words, providing the insured can prove there was a

case of COVID-19 at the premises prior to the Government

action, the policyholder can recover as the Government

action was proximately caused by the insured peril.

Prevention of access clauses

The Divisional Court in FCA v Arch (DC) found that

certain prevention of access wordings did not provide

cover. For example, clauses requiring:

• ”an incident… within a one mile radius… which results

in a denial or hindrance of access…”; or

• ”action by competent authority following a danger or

disturbance in the vicinity of the premises”; or

• ”action by competent authority following a danger or

disturbance in the vicinity of the premises”.

The Divisional Court held that the insured peril in those

clauses was only the local peril of the occurrences

of disease within the radius. By contrast, it found

that some prevention of access and other wordings

provided broader coverage and responded to any

wider or national action in response to the COVID-19

pandemic. The Supreme Court’s different approach

(ﬁnding that all radius clauses required a local peril,

but that proximate causation was satisﬁed even

though the “but for” test was not) opens the door to

undermining the Divisional Court’s ﬁndings on those

clauses, even though they were not appealed to the

Supreme Court. (The Supreme Court appears to have

indicated this at paragraph 250.)

The reasoning of the Supreme Court in FCA v Arch (SC)

opened the way for Lord Mance to express a preliminary

view in Policyholders v China Taiping Insurance (UK)

Co Ltd (10 September 2021) that the Supreme Court

causational approach applied to a prevention of access

clause providing cover for interruption or interference

in consequence of the actions of a competent local

authority due to “emergency threatening life or property

in the vicinity of the premises”, so that providing the

pandemic came within the vicinity there would be cover

in the same way as there was for the radius clauses in

the test case. (It was not necessary to express a deﬁnite

view on the issue because the policyholders’ claims

failed in any event as the relevant authority deﬁned

in the clause did not include the actions of a central

or countrywide authority such as the UK government.

For more information, see Legal update, COVID-19:

arbitrator dismisses business interruption insurance

claims.) This was followed by a ﬁnding by Cockerill J in

Corbin & King Limited and others v Axa Insurance UK Plc

[2022] EWHC 409 (Comm) that, based on the approach

to causation of the Supreme Court in FCA v Arch (SC),

a prevention of access clause providing cover for loss

resulting from interruption arising directly from the

actions of a statutory body in response to a “danger or

disturbance within 1 mile” also led to broad recovery

(that is for losses suffered as a result of restrictions

imposed by the government) provided there was a case

of COVID-19 within a one mile radius of the premises.

This decision has not been appealed. Some insurers in

London International Exhibition Centre plc v RSA and

others also accepted that a clause very similar to that

considered in the China Taiping arbitration but with the

words “immediate vicinity” would respond under the

FCA v Arch (SC) causation test save that there was a

disease exclusion in that clause.

A series of test cases behind the lead case, Gatwick

Investment Ltd v Liberty Mutual Insurance Europe SE, on

the applicable causation test for prevention of access

wording are being heard in autumn 2023.

13 Practical Law

Reproduced from Practical Law, with the permission of the publishers. For further information visit uk.practicallaw.thomsonreuters.com

Insurance contract law: causation

Causal links within the insured peril

As set out above, Within the insured peril, the composite

perils clauses (prevention of access and hybrid clauses)

considered in the FCA v Arch test case, and later China

Taiping, Corbin v King, London International Exhibition

Centre and Gatwick Investments, include causal links

within the perils themselves, for example, between the

occurrence of COVID-19 within 25 miles of the insured

premises and the public authority action. Indeed, that

was the link considered in the FCA v Arch test case, and

the same approach applies. The link between the disease

within the radius and authority action in a prevention

of access and hybrid clause had the same causal

requirement as that between disease and interruption in

disease clauses (paragraph 213, FCA v Arch (SC)).

The link between the harm and

loss and the measure of indemnity

In business interruption cover, the indemnity is for an

unliquidated amount of loss that was, for example,

caused by or resulted from, the relevant harm

(interruption) triggered by the peril (for example, ﬁre,

government action or disease). Accordingly, if there is

an insured peril within the meaning of the policy (for

example, ﬁre or disease), it must then be considered

whether the insured peril was the proximate cause of the

loss or damage. In the context of business interruption

policies, this is often explicitly recorded in trends

clauses, which are typically found in such cover and

which allow for adjustments to proﬁt ﬁgures to be made

to take account of any trends or circumstances which

would have affected the business even in the absence of

the insured peril (see below, Trends clauses).

There is a historical ﬁction that, at common law, an

insurer’s obligation under an indemnity insurance

contract gives rise to an action for unliquidated

damages, arising from the failure of the insurer to

prevent the insured from suffering damage, for example,

interruption or liability to a third party. Accordingly,

if there is an insured loss, the insurer is considered to

be in breach of contract and liable to pay damages

in the form of the insurance monies due under the

insurance contract (see Firma C-Trade SA v Newcastle

Protection and Indemnity Association Socony Mobil Oil

Inc and others v West of England Shipowners Mutual

Insurance Association (London) Ltd (No. 2) [1991] 2 AC 1,

per Lord Goff at page 35, who stated that “a promise of

indemnity is simply a promise to hold the indemniﬁed

person harmless against a speciﬁed loss or expense”).

The quantiﬁcation of the indemnity always depends

on the policy wording, but in general “the insured is

entitled to be put by the insurer into the same position in

which he would have been had the event not occurred,

but not a better position” (see Callaghan v Dominion

Insurance [1997] 2 Lloyd’s Rep 541 per Sir Peter Webster

at page 544 col. 2 and Endurance Corporate Capital

Ltd v Sartex Quilts and Textiles Ltd [2020] EWCA 308,

per Leggatt LJ (as he then was) at paragraphs 35-36).

Leggatt LJ’s analysis there was expressly based on the

“general object of an award of damages for breach

of contract”. (For more information on the decision

in Endurance, see Legal update, Property insurance:

Court of Appeal holds insured entitled to indemnity

on reinstatement basis even though insured had not

reinstated property.)

In FCA v Arch (SC) the Supreme Court resolved the

indemnity causation question (that is the causal link

between the harm or damage and the loss) in a similar

way to the question of the causal link between the insured

peril and the harm (interruption). Even where the policy

required that loss was “solely and directly” caused by the

interruption, it was held that the intention of the policy

was that other aspects of COVID-19 should not prevent

cover (for example, because people would have stayed at

home anyway). Accordingly, for the period during which,

and to the extent to which, the insured peril (for example,

disease and government action) is a proximate cause of

loss, determining the indemnity involves removing the

entire underlying fortuity (for example, COVID-19) and all

its effects, even though “but for” the insured peril some

of it would still have been suffered (for example, even

without closure of premises some people would have

stayed at home), and even if prior to the triggering of

the insured peril the revenue had dropped substantially

(for example, due to people staying at home and a

general downturn due to COVID-19). In essence, it was

held that it would be wrong to reduce the indemnity for

loss proximately by the insured peril, even if it was also

proximately caused by uninsured (but non-excluded)

perils with the same originating cause (paragraphs 228

and 294 to 296 of judgment).

For prevention of access and hybrid clauses this means

that for the purpose of determining the loss in the

context of COVID-19:

• For the period of prevention (or equivalent), all COVID-19

losses are removed (paragraphs 228 to 230).

• But this only applies to the part of the business for

which there is interruption, for example the in-person

part of a shop but not its web business, even if other

parts of the business are depressed by the effects of

COVID-19 (paragraphs 141, 283 to 286).

• Once that prevention has ceased, it is likely that only

losses proximately caused by the (former) prevention

will be recoverable. That is, losses solely caused by

the ongoing effects of COVID-19 on the community

will not be recoverable (paragraph 244 of FCA v

Arch (SC) and Hyper Trust Limited v FBD Insurance

14 Practical Law

Reproduced from Practical Law, with the permission of the publishers. For further information visit uk.practicallaw.thomsonreuters.com

Insurance contract law: causation

plc at paragraphs 215, 262-6). (See above, The effect

of the passage of time on proximate causation, for

consideration of the timing of causation.)

Trends clauses

Business interruption insurances typically incorporate

policy sections stipulating the methodology of quantifying

lost revenue, rather than leaving this to be determined

by the evidential and legal techniques of the general law

when quantifying claims for consequential loss in (for

example) contractual or tortious damages claims.

As explained in FCA v Arch (SC) (paragraph 253), such

provisions typically take the form of a formula under which

the loss is taken to be the amount by which the revenue

and gross proﬁt of the insured during the indemnity period

(usually 12 months) falls short of revenue (”standard

revenue”) and gross proﬁt earned in the equivalent period

a year or immediately before the date of the insured event,

as adjusted by a so-called “trends clause”.

The aim of a trends clause (usually expressly stated)

is to make such adjustments as necessary to the

calculation of gross proﬁt to provide for the trend of the

insured business, the intention being that the adjusted

ﬁgures will represent the results which would have been

obtained had the insured event not occurred.

A number of trends clauses of this type were considered

in the FCA v Arch test case, all of which referred to

the aim of the clause as being “to represent” “as

near as possible” or “as nearly as may be reasonably

practicable” “the results which would have been

achieved” “but for the damage” or “if the damage had

not occurred” (see FCA v Arch (SC), paragraph 256

(though the reference to “damage” is inappropriate in

business interruption cover which does not depend on

physical damage to the insured property, and should

better be understood in such cases as a reference to the

“insured peril”, see paragraph 257 of the judgment).

The proper approach to construction of such trends

clauses was explained in FCA v Arch (SC) at paragraphs

259 to 264, where three important points were made:

• First, trends clauses are part of the machinery for

quantifying loss, but do not deﬁne the scope of the

indemnity, that being the function of the insuring

clauses.

• Second, the trends clauses should if possible be

construed consistently with insuring clauses.

• Third, accordingly, trends clauses should if possible be

construed so as not to detract from the cover provided

by the insuring clauses, otherwise the quantiﬁcation

machinery is transformed into a form of exclusion.

Accordingly, the Supreme Court concluded that unless

the policy wording requires otherwise:

”the trends clauses should not be construed so as

to take away cover for losses prima facie covered

by the insuring clauses on the basis of concurrent

causes of those losses which do not prevent them

from being covered by the insuring clauses.”

(paragraph 264)

The Supreme Court held that the “simplest and most

straightforward way” to construe trends clauses

consistently with the insuring clauses was to recognise

that the aim of trends clauses “is to arrive at the

results that would have been achieved “but for” the

insured peril and circumstances arising out of the

same underlying or originating cause. Accordingly, the

trends or circumstances referred to in the clause for

which adjustments are to be made should generally

be construed as meaning trends or circumstances

unconnected with the insured peril” (paragraph 268).

Consistent with this, the court ultimately concluded that

trends clauses in the form considered by it should be

construed so that the standard turnover or gross proﬁt

derived from previous trading is adjusted only to reﬂect

circumstances which are unconnected with the insured

peril and not circumstances which are inextricably linked

with the insured peril in the sense that they have the same

underlying or originating cause. This was the reasoning

that was advanced but rejected in Orient Express Hotels

Ltd v Assicurazioni Generali SPA t/a Generali Global Risk

[2010] EWHC 1186 (Comm), a decision which has now

been overruled by the decision in FCA v Arch (SC) (see

also Riley on Business Interruption (Sweet & Maxwell, 11th

Edition, 2021) Chapter 4, paragraph 4.5).

Mitigation

In common law, damages for breach of contract and tort

are reduced to the extent that they were proximately

caused by the failure of the claimant to minimise the

loss. This is known as the duty to mitigate and is part

of the test for legal causation in contract and tort (see

BPE Solicitors v Hughes-Holland [2018] AC 599 (SC)

at paragraph 20, Lord Sumption. See also Practice

notes, Damages for breach of contract: an overview and

Damages in tort: an overview).

In marine insurance, insureds and their agents are under

a statutory duty to take reasonable measures for the

purpose of averting or minimising loss (section78(4)

of the Marine Insurance Act 1906). Marine policies

frequently contain a “sue and labour” clause, requiring

the insured to take reasonable measures to avert or

15 Practical Law

Reproduced from Practical Law, with the permission of the publishers. For further information visit uk.practicallaw.thomsonreuters.com

Insurance contract law: causation

minimise loss, and a corresponding obligation on

insurers to pay for the reasonable cost of such measures.

In other insurance, the insured is probably under a

common law duty to mitigate its loss following the

occurrence of an insured event (see MacGillivray on

insurance law, paragraph 31-008, City Tailors Ltd v Evans

(1921) 126 LT 439, page 443).

The question is not free from doubt, however, and

it appears arguable, at least under English law (in

contrast to that of Australia and New Zealand), that

the insured has no duty to mitigate its loss unless such

a duty is laid down in the policy (see Colinvaux’s Law of

Insurance (Sweet & Maxwell, 13th ed, 2022, Chapter 11,

paragraphs 11-299 and 11-300).

Perhaps as a result, many policies contain an express

mitigation condition. However, there are limits to what

the insured is expected to do at common law (that is

absent an express clause): so, for example, it has been

held, in the context of business interruption cover, that

the insured is not expected to take on fresh premises from

which to conduct its business; and, on the same footing,

proﬁts made at fresh premises should probably not be

taken into account in adjusting the claim (see City Tailors

per Lord Atkin at page 445. See also All Leisure Holidays

LT v Europaische Reiseversicherung AG [2011] EWHC 2629,

where in the context of travel insurance, the court held

that the insured was not under a duty to mitigate its loss

by taking another cruise offered by a different provider in

circumstances where its cruise had been cancelled). As

many policies contain wording addressing mitigation, it

is always necessary to consider the policy wording before

reaching a conclusion as to the extent of the insured’s

duty to mitigate in any particular case.

If the insured in good faith takes steps to prevent a

loss, and thereby creates another loss covered under

the policy this should be recoverable (as it would be

under common law contract and tort principles of legal

causation) (see, for example, Quinta Communications

SA v Warrington [2000] Lloyd’s Rep IR 81).

However, it may be different where the loss incurred

in preventing an insured loss is not itself loss of a kind

covered under the policy. For example, it was held

that sums paid by the insured water authority to carry

out work to avert or lessen its liability to a third party

were not sums it was legally liable to pay as damages

or compensation, and so were not recoverable under

the terms of the relevant insuring clauses in its liability

policy. The court doubted the proposition advanced

by the insured that the law requires an insured (in the

absence of express terms) to make reasonable efforts to

prevent or minimise loss which may fall to the insurers

and rejected the argument that there was an implied

term that the insurer would pay for such precautions

(see Yorkshire Water v Sun Alliance & London Insurance

Ltd [1997] 2 Lloyd’s Rep 21, per Stuart-Smith LJ at

paragraph 224).

The position appears to be different under property

insurance, where loss or damage caused by efforts

to avert or extinguish a ﬁre have been held to have

been proximately caused by the ﬁre so long as there

was reasonable justiﬁcation for the measures taken

(Stanleyv Western Insurance Co (1886) LR 3 Ex 71 and

Symington v Union Insurance Society of Canton Ltd

[1928] 12 WLUK 55). So long as the measures taken were

reasonable, it will probably be no defence for insurers to

show that no damage would have been suffered if those

measures had not been taken (MacGillivray on Insurance

Law, paragraph 26-012).

It has also been suggested that the doctrine of proximate

causation means it is unnecessary for the court to imply

a term in an insurance contract that an insured should

take reasonable precautions to avert or minimise loss.

The need to prove that the loss was proximately caused

by the insured peril leads to the result that if the insured

decides not to take a step which it would be reasonable

to take (or perhaps unreasonable not to take) its loss

will to that extent have been proximately caused by that

decision rather than the insured peril (MacGillivray on

Insurance Law, paragraph 26-014).

In summary a number of points should be borne in mind

when an issue arises as to the duty of the insured to take

steps to avert or minimise loss (or as to its entitlement to

recover the cost of doing so from the insurer):

• First, the terms of the policy must be checked, as an

express term may well provide the answer.

• Second, the answer will or may be different in different

classes of insurance such as liability and property.

• Third, the question whether an insured acts

reasonably or unreasonably will always be relevant,

if not necessarily determinative and causation may

provide an answer.

• Fourth, the duty of good faith may be engaged if a

decision (for example not to take action to avert a

loss) is taken in bad faith or, of course, fraudulently.

• Fifth, the authorities on this topic are at best difﬁcult

to reconcile and at worst conﬂicting, and ripe for

coherent review.

Aggregation

Insurance policies often include aggregation clauses

which provide for two or more events, losses or claims,

depending on the type of insurance cover, that are

covered by the policy to be treated as one event, loss

or claim where they are linked by a deﬁned connection.

Legal solutions from Thomson Reuters

Thomson Reuters is the world’s leading source of news and information

for professional markets. Our customers rely on us to deliver the

intelligence, technology and expertise they need to ﬁnd trusted answers.

The business has operated in more than 100 countries for more than

100 years. For more information, visit www.thomsonreuters.com

Insurance contract law: causation

Aggregation can beneﬁt the insured and the insurer: it

may, for example, enable an insurer to limit its liability;

or an insured to pool together its losses so that the

overall loss exceeds the policy excess or retention and

triggers the liability of the insurer.

The necessary connection between the events or claims

will often involve causational aspects. These may be very

general, for example, losses “arising out of one event”;

or they may be much more speciﬁc, for example losses

“directly attributable to one outbreak of disease”. In the

context of COVID-19 and business interruption, the latter

wording was said to be likely to give rise to “obvious

problems of what constitutes an outbreak and by what

criterion it is possible to judge whether a large number

of cases of a disease are all part of one outbreak or are

part of or constitute a number of different outbreaks”

(FCA v Arch (SC), paragraph 66)

The approach to interpreting aggregation clauses was

summarised by the Court of Appeal in Spire Healthcare

Ltd v Royal & Sun Alliance Insurance Ltd [2022] EWCA

Civ 17, considered in Legal update, Court of Appeal

considers “one source or original cause” aggregation

wording in liability insurance policy. In its judgment, the

Court of Appeal summarised a number of key points:

• The usual principles of contractual construction apply;

and such clauses are to be construed in a balanced

way without predisposition to narrow or broad

interpretation (per Andrew Baker LJ, with whom Bean

and Underhill LJJ agreed, at paragraph 20. See also

AIG Europe v Woodman [2017] UKSC 18). However,

some formulations have been held to achieve a broad

effect, for example “consequent on or attributable to

one source or original cause” (paragraph 21).

• There is in principle no distinction between an

“original” and an “originating” cause, both of which

connote a considerably looser causal connection

than “proximate cause” (paragraph 21 and Beazley

Underwriting Ltd v The Travelers Companies

Incorporated [2011] EWHC 1520). However, although

the original cause does not need to be the sole cause

of the insured’s liability (or loss), it is necessary to

ﬁnd a single unifying factor, and not every “but for”

cause is sufﬁcient to amount to an “original cause”

(paragraph 24).

It cannot be assumed that a particular word bears a

consistent meaning across all policies, because the

particular context of the same wording in different policies

(and even, occasionally, in different parts of a single policy)

may produce a different interpretation of that wording

(Colinvaux & Merkin’s Insurance Contract Law, paragraph

C-0161 and Midland Mainline Ltd v Commercial Union

Assurance Co Ltd [2004] Lloyd’s Rep IR 22).

For information on aggregation in insurance and

reinsurance, see Article, COVID-19: Aggregation in

insurance and reinsurance.

Key takeaway

Causation is a concept of fundamental

importance in many contexts in insurance and

can give rise to difﬁcult issues of fact and law.

Correct resolution of these issues depends not

only on the general law and the many decided

cases, but also, but often crucially, on the proper

interpretation of the relevant policy wording and

the application of business common sense. The

decision of the Supreme Court in FCA v Arch has

provided welcome certainty in some areas, for

example as to the analysis of concurrent causes

and the proper interpretation of trends clauses,

but further elucidation is to be expected from the

raft of business interruption cases now making

their way through the courts.