Swiss Re's attempt to clarify the definition of weather events.
An event, according to the Pocket Oxford English Dictionary, is “the occurrence of a thing, [or] the combined occurrence of two things.” By the standards of reinsurance wordings, the Oxford's lexicographers could not have been more vague. Small wonder, then, that judges presiding over reinsurance disputes have spent hundreds of hours deliberating over the intended meaning of such a seemingly simple little word, and have written reams of pages pronouncing what, exactly, constitutes an event.
Reinsurers constantly grapple with the issue. Committees in the London market have met repeatedly over the years to discuss ways to improve the so-called ‘hours clause', the industry-standard event definition that says a flood, for example, is one event if it lasts for less than 168 hours, and two if it lasts for longer. The debate may seem esoteric, but for insurers it is a particularly important distinction: if cover is exhausted by a catastrophic event, but the event exceeds the time limits of the relevant reinsurance cover, a reinstatement may well kick in, essentially providing a new, identical layer of cover for the same event. If reinsurers are strict about their aggregates, they would have to double them to take this double-hit factor into account.
On the other hand, the hours clause can mean that insurers face multiple retentions for the same event. It has been noted that even though the autumn floods in England last year were caused by three distinct cyclonic depressions, strict interpretation of hours clauses could force some insurers to make six reinsurance recoveries (provided they had sufficient reinstatements), and thus to bear six separate retentions for what climatologists might argue were only three actual weather events. If the latter was recognised, reinsurers would pay an amount equal to three deductibles, but the likelihood that their cover is exhausted would also increase.
Swiss Re has embarked upon a major effort to address the problems. From the last renewal, the world's largest reinsurer has adopted a new, home-made event clause based more on science and less on time. The wording describes as ‘single events' a “storm due to an atmospheric disturbance usually so designated by a meteorological institute; hail and/or thunderstorms and/or tornadoes due to an atmospheric disturbance; earthquake, tsunami, volcanic eruption, flood by one and the same instance of high water which may have more than one peak and which may occur in one or more bodies of water; conflagration;” and “strike, riot, civil commotion or violent demonstration occurring within the boundaries of one city, town or village.”
“The intention is very similar indeed to most good, traditional hours clauses,” said Nikki Fritz, product specialist at Swiss Re. “What we did was reconsidered the element of time.” The change should eliminate the possibility of multiple coverage of single events. “Within the text of the new event definition there is no provision for reinstatements. That is a result of the logic that the cover is for one event,” Ms Fritz said. “The potential for reinstatements in a traditional hours clause is both positive and negative. We have to respect the feeling that some ceding companies will view this as a kind of back-up, but it would be much better for them in the first place to do an accurate assessment of their exposure, and purchase adequate reinsurance.”
Acceptance of the new event clause has not been widespread, but it has found some takers. “We have been successful in introducing it in a number of cases,” said Gordon Fox, a property reinsurance underwriter at Swiss Re UK. “Some small UK companies and a number of syndicates have it throughout their programme.” In addition, one Lloyd's syndicate has adopted the clause for all its inward reinsurance business. Only one cedant has refused the clause outright, Mr Fox said, while some larger insurance groups have asked for more detail. “We are quite pleased with the way it has been adopted,” he said, adding that Swiss Re would not insist upon the clause, and pointing out that cedants should have the same event definitions throughout their programmes to avoid coverage gaps.
However, the reinsurance buyer for a major European insurance group reported that he did not adopt the clause at the last renewal. “The science is only theoretical. In terms of weather, there is still a big question mark about what is an event,” he said, pointing to the storms in France in 1999, the same storms that Swiss Re cites as an example in its explanatory note to the event clause. “According to some science, Lothar and Martin were caused by a single event called Kurt, which spun off Lothar and Martin,” the cedant said, “but Swiss Re argues that a single event does not go this far. They say it is the event that causes the damage that is the event.”
He finds the argument flawed. “You are either scientific or pragmatic. This approach seems to be having your cake and eating it.”
Swiss Re's explanatory notes read as follows: “Although there was some initial discussion as to whether or not Kurt, the large scale weather system off the coast of Iceland which ‘fathered' Lothar and Martin, should be considered an event for reinsurance indemnification purposes, common sense soon prevailed.”
Chain of causation
Clearly, what seems like common sense to Swiss Re is not so obvious to the reinsurance buyer, but Swiss Re's argument goes no further, and it appears strong: the chain of causation begins with the event that brings losses. “Kurt didn't cause any insured damage,” Mr Fritz pointed out. “It was caused by Martin and Lothar.” The two storm events originated from separate and distinct storm fronts, took place on different dates, and followed different paths, which did partially overlap, Swiss Re stated.
“We, as buyers, are looking for protection not from the cause, but from the effect it has on us, and the hours clause tries to do that,” the reinsurance buyer said. “The current hours clause has got some weaknesses, yet one of the problems with events is that it is quite hard to define one until it happens. You don't know it until you see it.” He drew a distinction between the certainty of the science-based event definitions which are required for the transactional-based catastrophe risk transfer of securitisation and e-commerce, versus the relationship factor of conventional reinsurance trading, where less certainty is required and where the understanding that reinsurance protects cedants from the effect of the cause prevails.
“It comes down to something fundamental,” he said. “Cedants need a contract which allows for the uncertainty inherent in the nature of catastrophe losses. In some situations you have to be pragmatic.” Mr Fox argued that the clause does just that, but from a scientific basis. “Our view is that the recovery should attempt to match the loss, instead of something artificial,” he said.
When science cannot define the event, the Swiss Re wording reverts to an hours clause. However, it recognises that disagreement over the scientific approach may arise without defaulting to hours. “There is obvious potential for uncertainty or differences of opinion as to how many events have indeed occurred,” Swiss Re points out in its explanatory notes. The wording states that: “In case of uncertainty over scientific issues, the parties agree to seek expert advice from a neutral and recognised organisation,” a clause intended “to settle such questions amicably on the basis of scientific fact outside of court and prior to formal arbitration procedures whenever possible,” the notes proclaim.
Ms Fritz said it is the intention that disagreements over numbers of events could be put to one of the handful of international institutions with expertise in the area. “If push really came to shove, and there is genuine disagreement, you have the arbitration clause in the treaty, but only in the rarest cases should it go that far.” Colleague Mr Fox said Swiss Re would listen, should clients wish to name an expert arbitrator in the contract. “I would be surprised if we came across a situation where we did not find agreement with the client,” he said, adding that it is common practice to refer to a weather bureau for confirmation of events under an hours clause.
Although many cedants have not embraced the wording, it has been broadly welcomed. Swiss Re's sales pitch, one broker reports, mentions that the possibility of mid-event reinstatements means a higher capital charge is included in reinsurance pricing, hinting that the current rates reflects the possibility, and thus that the new clause should lead to lower prices, or at least to less volatility. And despite his rejection of the clause, the cedant is supportive of the principle of revisiting events definitions. “We do need to have a debate about this,” he said. “We need to clarify the issues from both buyers' and sellers' perspectives. Swiss Re's method has been to put a proposal forward, and now we need to agree what the issues are, and see if we can find a solution.” A paper explaining the application of the clause to the UK floods of 2000 was due, at the time of writing, to be released shortly. It will help the process along by furthering the debate. Meanwhile Swiss Re is patient, seeing the introduction of the clause as only the beginning. “We understand that it will take time,” Ms Fritz said. “Things don't change overnight.”
Swiss Re's event definition for property reinsurance
1) For the purposes of this reinsurance agreement, an event shall include all insured losses which arise directly from the same cause and which occur during the same period of time and in the same area. Such cause is understood to be the peril which directly occasions the losses or where there are several perils which, in an unbroken chain of causation, have occasioned the losses, the peril which triggered the chain of causation.
For example, as long as they are covered by this reinsurance agreement, losses occasioned by the perils set out below at letters a) to f) shall constitute single events:
a) storm due to an atmospheric disturbance usually so designated by a meteorological institute;b) hail and/or thunderstorms and/or tornadoes due to an atmospheric disturbance;c) earthquake, tsunami, volcanic eruption;d) flood by one and the same instance of high water which may have more than one peak and which may occur in one or more bodies of water;e) conflagration;f) strike, riot, civil commotion or violent demonstration occurring within the boundaries of one city, town or village.
2) If the number of events cannot be determined according to paragraph 1, the following hours clause is then applied. An event shall thus encompass a continuous period of time starting with the occurrence of the reinsured's first individual loss and lasting
In the case of differing perils which are not connected to each other by an unbroken chain of causation, the applicable number of hours corresponds to those of the peril which has caused the largest amount of damages.
3) In the case of more than one event, if it is impossible to allocate any losses, the reinsured shall allocate them to the event whose cause is most likely to have occasioned them.
In case of uncertainty over scientific issues, the parties agree to seek expert advice from a neutral and recognised organisation.