After Japan: the intricacies of semantics


John Wasty

After Japan: the intricacies of semantics

Events in Japan have once more raised the significance of reinsurance clauses as the industry grapples with the causality of contracts.

The recent natural disasters in New Zealand, Japan and the United States resulted in thousands of deaths and produced losses for catastrophe reinsurers estimated at between $20 and $45 billion. Reinsurers in Bermuda and other international reinsurance centres will have significant exposures to these losses. This article outlines some of the reinsurance contract wording issues likely to face reinsurers and their reinsureds.

One concern that arises in relation to the Tohoku earthquake, and the tsunami caused by it, is the issue of concurrent losses. Where two causes of loss appear to operate concurrently, the question of whether or not a loss is covered hinges on whether each cause led to independent loss. As with the earthquake and tsunami, a subsequent cause may join with a previous and continuing cause, with the result that the two causes become concurrent some time after the initial cause began.

Two key issues for reinsurers in these situations are firstly, aggregation of losses arising from these events and secondly, hours clauses.


For catastrophe excess of loss reinsurers, aggregation is frequently a threshold issue. Excess of loss contracts typically are triggered per event or per occurrence. Where there are arguments that losses were caused by more than one event, the issue of causation is also likely to be a key consideration for reinsurers and their reinsureds.

In two seminal cases, the English courts have sought to define what constitutes an ‘event’. In Axa Reinsurance (UK) v Field 1996, Lord Mustill held:

“…an event is something that happens at a particular time, at a particular place and in a particular way.”

while a “cause” is:

“….something to my mind altogether less restricted. It can be a continuing state of affairs; it can be the absence of something happening.”

The same year, in Kuwait Airways Corp. v Kuwait Insurance Co SAK [1996], the court approved a ‘unities’ test for defining an ‘event’ that involved assessing the following four factors from the viewpoint of an informed observer: an event is characterised by unity of cause, locality, time and—where human action was involved—intention.

Having identified a potential event using these principles, reinsurers and their reinsureds will then consider whether or not certain losses are capable of aggregation, i.e whether two or more separate losses covered by the policy may be treated as a single loss such that the policy deductible is applied to the cumulative (aggregate) loss, rather than being applied individually to each loss under the policy.

A clause covering losses “arising from one event” refers to losses causally connected with that event, notwithstanding it need not have been the proximate or immediate cause. Nevertheless, that event cannot be too remote from the loss suffered and the causative link must be significant. Conversely, where clauses provide that losses must be a “direct and immediate result of” or “directly occasioned by” an event, the losses must be more proximal to the catastrophe or insured peril that is reinsured.

The Japanese earthquake allows an examination of a complex series of events in relation to a chain of causation: a severe offshore underwater earthquake gave rise to a tsunami, which destroyed vast portions of Japan’s north-eastern coast. Following the earthquake and tsunami, Japan suffered significant aftershocks, massive fires, a nuclear power plant meltdown and power losses, and the value of some of Japan’s largest companies fell sharply.

Reinsurers in these circumstances will examine the chain of events in order to assess whether there was a break in the causation chain between the original insured peril and the loss claimed. This is recognised by arbitrators and courts as a thorny and fact-intensive issue that can again give rise to disputes: how may damage from the tsunami or the aftershocks be distinguished from damage caused by the earthquake, and in what circumstances should it be?

Hours clauses

Reinsurance contracts may also contain ‘hours clauses’ in place of more familiar aggregation clauses. These vary from ‘event-based’ to ‘manifestation of insured peril’ to ‘catastrophe-based’ wordings and are often subject to bespoke amendments. As there are no Bermuda or English cases on the application of hours clauses, they are construed according to basic tenants of contract interpretation and such legal principles as causation. The key elements for any hours clause— the definition of loss occurrence—is as varied as the potential for aggregation clauses. The event-based clause (NMA 2244) covers losses directly occasioned by a series of disasters or accidents that arise out of a single event and is at the broader end of the spectrum.

An hours clause then deems a ‘cut-off’ for the period of the loss occurrence up to which losses may be claimed: typically fixed at 72 or 168 hours, contingent on the peril. In the case of the former, reinsurers will generally grant the reinsured a further 72-hour period that may be treated as a separate ‘loss occurrence’. In that light, the manner in which an ‘event’ is defined may give rise to disputes. A clause covering all losses that are ‘directly occasioned’ by, say, an earthquake and happening over the following 72-hour period may allow aggregation by the reinsured, but is the reinsurer also able to include losses arising from the tsunami?

Additional issues arising out of the series of disasters in Japan include whether property damage was proximately caused by the earthquake or the subsequent tsunami. If the latter, the reinsured will want to argue that the tsunami was part of the earthquake in order to maximise its recovery, i.e. the proximity between the earthquake and tsunami was so great that it should actually be considered part of the earthquake for the purposes of the hours clause trigger and thus a single loss occurrence. Reinsurers will consider whether such losses may and should be separated. Furthermore, disputes may arise if it appears that more than one loss occurrence or catastrophe has arisen regarding how specific losses are to be allocated to particular parts of the hours clause.

Industry loss warranties/original loss warranties

Certain catastrophe and excess of loss insurance contracts (industry loss warranties—ILWs) are only triggered if a market loss exceeds a designated threshold. The preceding discussion provides a flavour of the complexities of assessing earthquake and tsunami losses in Japan given the practical challenges associated in assessing, say, long-term business interruption—which may itself arise from multiple causes—and the significant structural damage to property. In the circumstances, it may be some time before market losses can be quantified to determine whether ILW contracts have been triggered. If the ILW definitions are themselves bespoke, issues may also arise as to how they are to be applied given the complexities of the losses.

In Bermuda, major reinsurers have availed themselves of catastrophe bonds that are placed with special purpose vehicles (SPVs) supported by note-holders. While catastrophe bonds offer conventional indemnity cover, they commonly operate on a parametric basis such that on the occurrence of a specified event of a certain magnitude, e.g. an earthquake of a certain magnitude in a certain region, then a predetermined benefit becomes payable. In such instances, the trigger may give rise to an immediate payout, although the benefits under the contract may also take the form of additional coverage for subsequent events in the applicable period of cover.

Typically, catastrophe bonds are carefully worded and should provide relatively little scope for dispute between their holders and the SPVs. That said, the confluence of events in Japan, and the difficulty in collating the evidence required to determine whether these contracts have been triggered, may delay payouts.

Nuclear exclusions

Underlying policies usually contain a nuclear exclusion, which would be mirrored in reinsurance policies. Damage arising from radiation sickness, losses from the exclusion zone and irradiation of food and water may, accordingly, be excluded. That said, by virtue of the Japanese Financial Supervisory Agency’s approval of all domestic wordings, such exclusions may be ineffective, which may result in disputes between reinsureds and their reinsurers.

The market

The significant loss estimates of these disasters have given rise to the suggestion that the reinsurance market might harden. If so, the usual bargain will need to be struck between the reinsured negotiating for clauses that cover the types of losses they anticipate, and reinsurers being in a position to demand sufficient premium to cover such eventualities. Conversely if, despite the magnitude of the losses, the market does not harden, reinsureds will have more scope to negotiate clauses in their favour, at lower, or at least similar, rates. The dynamic need not be a hostile one given that both sides merely seek to obtain what they want at an acceptable price.

Arbitration, and litigation, will inevitably arise concerning these issues. Given the confidential nature of arbitration, it may be that resolutions of some of the issues identified above are not publicised. That said, the results of such arbitrations do eventually feed into either demands made by reinsureds to cover such eventualities if the arbitration is decided against them or result in reinsurers taking steps to address any wording loopholes exposed by the results of arbitration proceedings.

John Wasty is group team leader, insurance and reinsurance dispute resolution at Appleby. He can be contacted at:

Japan, Tohoku, reinsurance, clauses, earthquake

Bermuda Re