Construction Risks

Archer v Ace (or, The Demise Of LEG3?)

Introduction

In the London Market there is, by and large, a common understanding about how LEG3 and the other defects exclusions operate, and what they are intended to do. That doesn’t mean that disagreements don’t arise about how a particular defects exclusion might apply to a particular set of facts, but those disagreements tend to be relatively rare, and the London Market tends to deal with what we call Construction All Risks claims (or what would be known in the US as Builders’ Risk claims), quite well on the whole.

As a result, those using the defects exclusions in the London Market, whether that is insurers, brokers, or the more sophisticated policyholders, tend to overlook the fact that several of the clauses, in both the LEG and DE suites of exclusions, are actually very difficult to understand for those who come to the clauses with the (surely reasonable) aspiration of wanting to determine the meaning of the clauses from the words that they contain.

Towards the end of last year I wrote about the potential impact of the first Court decision anywhere in the world which considered the meaning of the defects exclusion which (along with DE5, which is much less commonly used outside the UK) is intended to preserve the most generous coverage for damage to works under construction, LEG3, in the case of South Capitol Bridgebuilders v Lexington. That case was decided by a Court in the District of Columbia, but applied the Law of Illinois. Now, like buses, a second decision has been handed down in the US which considers LEG3, this time applying the Law of Florida, in the case Archer Western - De Moya Joint Venture v Ace American Insurance Co.

The decision in SCB sounded alarm bells for the Builder’s Risk community in the US, and presumably also for the LEG Committee, who are responsible for the LEG defects exclusion clauses. It raised at least two questions of significance: what constitutes damage for the purpose of triggering a Builder’s Risk policy; and what is the meaning of the LEG3 clause? Its answers to those questions were striking: property that from an English law perspective would have certainly been regarded as merely being in a defective condition was held by the Court in SCB to have suffered damage. With regard to the meaning of LEG3, the Court in SCB appeared to be unable to form a view, and held that the clause was ambiguous: “egregiously so”.

The big question for those who, like me, have an interest in the health of the Builder’s Risk market, was whether SCB would come to be regarded as an outlier decision, or one that would have a meaningful impact? Archer v Ace suggests the latter.

The judgment in Archer concerns an application for summary judgment by the insurer which was denied, and so the issues in the case will continue towards a substantive trial in due course. However, the judgment runs to some 66 pages, and so the issues were considered in some detail. I am not going to try to cover all of the detail but, as with my article on SCB, am going to focus instead of what the most important elements might mean for the Builder’s Risk market.

The facts

Again, I’ll start with a very brief description of the facts, which up to a point may create a sense of Deja Vu for those familiar with the SCB decision. Once more we have a Builder’s Risk claim relating to inadequate concrete in a bridge under construction. We have a disagreement about whether the works under construction were damaged (so as to trigger the Builder’s Risk policy), or whether the works were merely defective (which would not trigger the policy). We have a policy that contains a LEG3 defects exclusion. And we have disagreements about what LEG3 means, and about how one might establish what constitutes the “improvements” with which LEG3 is concerned.

In Archer the policyholder was a design and build contractor for the snappily titled “I-395/SR 836 Reconstruction / Rehabilitation Project” in Miami, Florida, which included the construction of a “signature bridge”. The design of the bridge involved batches of concrete, the production of which included the addition of “fly ash” from a pressurised fly ash silo, which had a mechanical system which was intended to allow specified amounts of fly ash to be added to the concrete batches. At some point between August and November 2020 the pressure relief valve of the silo failed, so that certain batches were “adulterated by an excessive amount of fly ash”.

I am not my firm’s expert on concrete (the “I ❤️ concrete” mug on my colleague Joanna Grant’s desk probably tells you who is) but, as the Court explained in Archer, although cement and concrete are terms that are often used interchangeably, they aren’t the same. Rather, cement is one of the ingredients of concrete, with the other common ingredients of concrete being fly ash, water, and aggregates. So, the presence of fly ash in concrete is not a problem in and of itself. In fact, in one sense, the more fly ash there is in the concrete, the better, as long as using additional amounts of fly ash does not come at the expense of the amount of cement used. High proportions of both fly ash and cement “generally increases the overall compressive strength of the concrete”. The problem comes when, as in Archer, additional amounts of fly ash are used at the expense of the amount of cement used. Then the compressive strength of the concrete is impaired.

When the policyholder became aware that some of the concrete had inadequate compressive strength, it submitted a claim for indemnity for the cost of repairing the concrete. The insurer denied coverage “reasoning the concrete constituted a defective material due to to the excess fly ash, and `because of this defect the material was never in a satisfactory state and therefore was not damaged’”.

Based on the above, the Court was required to address the following questions:

• Did the insured property suffer damage?
• Is LEG3 ambiguous?

In approaching those questions, the Court applied the test for summary judgment under the Law of Florida, which is that “summary judgment is appropriate where there is ‘no genuine issue as to any material fact’, and the moving party is ‘entitled to judgment as a matter of law’” (per Federal Rule of Civil Procedure 56), and that “when deciding whether summary judgment is appropriate, the court views all facts and resolves all doubts in favour of the non-moving party”.

It also applied the test for ambiguity under the Law of Florida, which is that “a policy is ambiguous only when ‘its terms make the contract susceptible to different reasonable interpretations, one resulting in coverage and one resulting in exclusion’”, and that “if there is an ambiguity, then it is construed against the insurer and in favour of coverage”.

As I did in my SCB article, I’ll explain what the Court held in relation to each issue, and add some comments of my own.

Did the insured property suffer damage?

As with SCB, the policy in Archer didn’t define the term “damage”. However, rather than just going to the dictionary, as the judge had done in SCB, the judge in Archer held that the test for damage had been determined by previous cases, and that it “requires a tangible alteration to the covered property”. That test is largely consistent with the test under English law, which requires a change in the physical condition of the insured property, which impairs the value or the usefulness of that property.

On the facts, and based on the high bar required to give summary requirement, the judge was “not prepared to accept the insurer’s argument that damage to the cement did not involve a physical alteration” and so that issue will remain to be determined at trial.

From an English law perspective, the issue is an interesting one, and the correct answer is not obvious. The correct answer will, in my view, turn on what is considered to be the relevant property: the concrete, or the cement?

If I was representing the policyholder, I would be arguing that the relevant property is the cement, and that the cement has become damaged by being overlaid with excessive quantities of fly ash. We know, from cases such as Hunter v Canary Wharf and R v Henderson, that the deposit onto insured property of excessive quantities of benign substances is capable of constituting damage, where the excessive quantities of those substances cost more money to remove than if ordinary quantities of those substances were present. On that basis, I would argue that the cement has undergone an adverse change in physical condition, that impairs both its value and its usefulness by coming into contact with excessive amounts of fly ash: the policyholder started out with cement which had a particular value, and as a result of the change in physical condition that occurred when the fly ash was added, it no longer retains that value.

If, on the other hand, I was representing the insurer, I would be arguing that the relevant property is the concrete, and that it was in a defective condition from the moment it was created (by the mixing of the cement and the fly ash). I would argue that from that point onwards it didn’t undergo any further “tangible alteration”, meaning that the test for damage hasn’t been satisfied. We know from the Bacardi case that, in English law, the creation of a defective finished product doesn’t constitute damage. Although Tioxide tells us that damage does occur when a defective finished product undergoes a change in physical condition that constitutes a further impairment of value or usefulness, that hasn’t happened in Archer, where the concrete was under-strength as soon as it came into existence, and remained that way until discovery.

So, which material should the Court be concerning itself with, the cement or the concrete? Although, as a policyholder representative, I would like to say that the Court should be concerning itself with the cement, I don’t think that’s right. The property which needs fixing is the concrete. The claim is not for the cost of repairing the cement, but for the cost of repairing the concrete.

On the basis of the above, although the insurer wasn’t successful in obtaining summary judgment on the proposition that the insured property hadn’t suffered damage, I expect the insurer to succeed on that issue at trial.

Is LEG3 ambiguous?

As in SCB, the Court in Archer first considered whether it was ambiguous as to whether LEG3 was an extension or an exclusion. The policyholder had argued that LEG3 is “both a coverage grant and an exclusion”, and the Court held that LEG3 “generates a functional extension, or broadening, of coverage”, as compared with the narrower exclusion which LEG3 had replaced by endorsement.

That doesn’t sound right to me, and in my view that doesn’t reflect the position under English Law. Tesco v Constable makes clear that the main insuring clause of a policy can only be widened by other clauses in the policy by using the clearest terms (and ABN Amro then gave an illustration of just how clear those terms needed to be, i.e very).

The second potential ambiguity in LEG3 was what it means to “‘improve’ the original workmanship”. Here, the Court in Archer didn’t develop the arguments any further than in SCB, and simply agreed that LEG3 was ambiguous in that regard.

So, where does that leave us?

In a few short months two different Courts, applying the law of two different States, have both held that LEG3 is ambiguous. In fact that’s being somewhat diplomatic, and it’s probably more true to say that neither Court could work out what on earth LEG3 was supposed to mean. That being the case, if SCB suggested that there was an opportunity for the LEG Committee to take a fresh look at the drafting of LEG3 and the other defects exclusions, Archer suggests that it really has no option, and that it must do so as a matter of urgency.

If LEG3 is going to be amended (as, in my view, it must), then the LEG Committee also has an opportunity to overhaul the other defects exclusions.

Although the DE clauses and the LEG clauses have different origins, it is not helpful for there to be two different suites of clauses which are so similar to each other. In my view it would be much better for there to be a single suite of clauses which captures the best elements of the current clauses.

So:

• There should be a clause which is concerned with causation, and which excludes the cost of repairing any damage caused by mistakes (which would essentially be a re-drafted, simplified, version of DE1 and LEG1, which both do the same thing);
• There should then be two clauses which are concerned with the condition of the relevant property before the damage occurs. One of those clauses would exclude the cost that would have been incurred to repair any defects which were present in property that has become damaged, if those defects had been discovered immediately before the damage occurred (i.e. a re-drafted, simplified, version of LEG2). The other clause would exclude entirely the cost of fixing damage to property which was in a defective condition immediately before the damage occurred (i.e. a re-drafted, simplified, version of DE3, which one might call LEG2A in the new suite);
• The final clause would exclude only the cost of improvements (i.e. a re-drafted, simplified, version of LEG3). My SCB article proposed an amended version of LEG3, and a few months later I would still stand behind that draft.

Those clauses would be made to be bought together. So, a policy with the most limited cover would contain only LEG1. A policy with wider cover would contain both LEG1, and also either LEG2 or LEG2A (whichever is most appropriate for the type of project involved). A policy with the widest cover would contain LEG1, plus one of LEG2 or LEG2A, and also LEG3. Where a policy contains more than one of the new defects exclusions, the policyholder should be able to choose which to apply in the event of a claim, with each exclusion coming with a different deductible. LEG1 would have the lowest deductible. LEG2 or LEG2A would have a higher deductible, and LEG3 would have the highest deductible of all.

That, in my view, would represent a very healthy outcome for insurers, brokers, and policyholders alike, and constitute a positive response to the issues raised by SCB and Archer: a single suite of defects exclusions; which are simply drafted and easy to understand; and which fit together with each other, and are intended to be used in conjunction with each other.

David Pryce is the Managing Partner at Fenchurch Law

by David Pryce

Covid “Catastrophe” Triggers BI Reinsurance

The first UK court ruling on the reinsurance of Covid-19 losses has confirmed coverage under excess of loss policies taken out by Covéa Insurance plc (“Covéa”) and Markel International Insurance Co Ltd (“Markel”). Mr Justice Foxton allowed recovery against reinsurers for losses occurring while the underlying policyholders were unable to use their business premises, due to government restrictions, on the basis that the pandemic was a “catastrophe” within the meaning of the reinsurance contracts.

Covéa and Markel paid out a combined total of over £100 million to policyholders for Covid-19 business interruption (“BI”) losses and made claims under their respective reinsurances with UnipolRe Designated Activity Company and General Reinsurance AG. Disagreements arose concerning the scope of cover under the reinsurance contracts, and a consolidated judgment was given in two separate appeals under s.69 of the Arbitration Act 1996, against arbitration awards dated January and July 2023. In summary, the appeals raised the following issues:

1. Whether the relevant Covid-19 losses arose out of and were directly occasioned by one catastrophe on the proper construction of the reinsurances. Both arbitration awards found that they did; and

2. Whether the respective Hours Clauses in the reinsurances, which confined the right to indemnity to “individual losses” within a set period, meant that the reinsurances only responded to payment in respect of the closure of insured premises during the stipulated period. The Markel arbitral tribunal found that the relevant provision did have that effect, while the tribunal in the Covéa arbitration found that it did not.

The Judge found in favour of the reinsureds on both issues.

Loss Arising from One Catastrophe

Coincidentally, both Covéa’s and Markel’s losses arose through direct insurance of nurseries and childcare facilities, which had been forced to close from 20 March 2020 by the UK government’s Order of 18 March 2020. The reinsurance contracts contained similarly worded Hours Clauses based on the LPO 98 market wording, including a form of aggregation provision operating by reference to a specified number of hours’ cover for any “Event” or “Loss Occurrence” (terms previously held to have the same meaning), defined as “all individual losses arising out of and directly occasioned by one catastrophe”.

For any Event or Loss Occurrence “of whatsoever nature” which did not include losses arising from specified perils (such as hurricane, earthquake, riot or flood) listed in the Hours Clauses, the limit was 168 hours (i.e. 7 days). Infectious disease was not a named peril and the 168 hours limit applied.

In circumstances where the arbitral awards were based on mixed findings of fact and law, it was common ground that the court could not interfere on a s.69 appeal unless the arbitral tribunal either had erred in law or, correctly applying the relevant law, had reached a decision on the facts which no reasonable person could have done.

On general principles of construction, the Judge endorsed the comments of Mr Justice Butcher in Stonegate v MS Amlin [2022] that “in considering whether there has been a relevant ‘occurrence’, the matter is to be scrutinised from the point of view of an informed observer placed in the position of the insured” (per Rix J. in Kuwait Airways [1996]).

Reinsurers argued that the gradual unfolding of the pandemic did not qualify as a “catastrophe” under the reinsurance policies, taking account of the historical development of property excess of loss market wordings, which was said to implicitly demonstrate the requirement for a sudden and violent event or happening, which could not be established on the underlying facts. Further, reinsurers claimed that a catastrophe is a species of occurrence or event that must satisfy the “unities” of time, manner and place, applied by Lord Mustill in Axa v Field [1996].

The Judge concluded that terms of the reinsurance contracts supported a generous application of the unities test, given the requirements for losses under multiple policies, with a duration potentially exceeding 504 hours (the period specified in relation to flood perils, i.e. 21 days), within broad geographical limits, indicating that a covered catastrophe could have a potentially wide field of impact.

While acknowledging the difficulties inherent in distinguishing between a “catastrophe” properly so-called, as an appropriate basis for aggregation, and a series of discrete losses sharing some common point of ancestry, the Judge held for purposes of the reinsurance claims under consideration that a catastrophe:

• Must be capable of directly causing individual losses, likely in most cases to exclude “states of affairs”.
• Is a coherent, particular and readily identifiable happening, with an existence, identity and “catastrophic” character arising from more than the mere fact that substantial losses have occurred.
• Will be identifiable, in a broad sense, as to its time of coming into existence and of ceasing in effect.
• Involves an adverse change on a significant scale from that which proceeded it.

Applying these principles to the findings in each award, the Judge noted that both tribunals had referred to the outbreak of Covid-19, and the resulting disruption of life in the UK, leading up to and necessitating the 18 March Order, as a catastrophe. In circumstances where the various government directives, including the 18 March Order, were rational and considered measures taken in the public interest, it was not necessary to explore the issue of whether a government order in isolation could be viewed as a catastrophe, since “the pandemic and the response thereto could not be disentangled”, an approach consistent with the decisions in Star Entertainment v Chubb Insurance Australia [2021] and Gatwick Investment v Liberty Mutual [2024].

Interpretation of the Hours Clauses

Covéa and Markel argued that all BI losses arising from the 18 March Order were reinsured, notwithstanding the BI losses continuing after expiry of the 168 hour period. The Covéa arbitral award endorsed this approach, determining that the reference to “individual loss” meant “a loss sustained by an original insured which occurs as and when a covered peril strikes or affects insured premises or property”. However, the Markel tribunal found in favour of the reinsurer, reasoning that it was “natural to think that BI losses occur day by day”, and therefore construing the relevant words as not “dealing with causation but with the occurrence of a particular loss”, since the “subject matter of an ‘Event’, its duration and extent, and its occurrence, are all referenced to losses not perils.”       

The effect of the Markel tribunal’s finding was that only 168 hours of BI losses could be recovered from the reinsurer (although the BI had in fact extended until at least June 2020, when the relevant restrictions were first lifted), so that most losses fell outside the scope of cover and Markel was unable to reach the specified attachment point under the reinsurance policy.

In reconsidering this issue on appeal, the Judge was not persuaded that a clear line could be drawn between damage and non-damage BI, as contended by reinsurers, since even the former might continue to manifest after the specified hours period, for example by damage worsening over time. Further, the Court concluded that the “wait and see” analysis applied by the Markel tribunal, premised on the occurrence of BI losses on a day-by-day basis, may lead to “uncommercial consequences” and does not sit easily with the findings in Stonegate and Various Eateries v Allianz [2022], which treated the closure orders as having immediate impact on the insured property with continuous effect, analogous to physical damage to buildings; or with the Supreme Court’s decision in FCA v Arch [2021], which suggested that the correct causal sequence for non-damage BI approximates that of damage-related BI.

The Court therefore dismissed the reinsurers’ appeals as to the meaning of “catastrophe” and allowed Markel’s appeal against the conclusion of the arbitral tribunal as to the effect of the Hours Clause.

Practical Implications

Figures published by the Financial Conduct Authority in March 2023 indicate that, since conclusion of the Test Case in 2021, insurers have paid around £1.4 billion in BI claims. The Commercial Court’s decision in this case provides comfort for cedants with ongoing recoveries, significantly restricting reinsurers’ ability to challenge the presentation of Covid-19 losses under similarly worded excess of loss property policies. It will be interesting to see how the decision may be applied in subsequent cases involving aggregation of losses across multiple jurisdictions. Given that reinsurance contracts typically provide for resolution of disputes by way of (confidential) arbitration proceedings, this clear judgment in favour of the cedants is particularly illuminating.

UnipolSai Assicurazioni SPA v Covea Insurance PLC [2024] EWHC 253

Amy Lacey is a Partner at Fenchurch Law

by Amy Lacey