Holding the Key to Climate Change

Common-sense underwriting and current research on climate change are useful supplements for existing deficiencies in Canada’s wind and weather catastrophe models, according to panelists at a Swiss Re seminar held in Toronto.

Reinsurers would be well-advised to compare their actual experiences to modeled events when underwriting catastrophe exposures, according to seminar panelist Sean Russell, the head of Swiss Re Canada’s underwriting unit.

“So how good have these models been?” Russell asked his audience. “If you look at what’s happened in the past 10 years, how does that correlate with what the models are producing?

“If you’ve had five, 1-in-100-year events in the past 10 years, then they’re not 1-in-100-year events. The models are probably not that accurate.”

Andrew Castaldi, the head of Swiss Re’s cat perils unit, said he is not a “fan” of models applied to the Canadian market. “I am a fan of accumulations, of managing your exposures. The models for these kinds of events are always going to be off.”

Castaldi said after the 2005 hurricanes, reinsurers pegged an acceptable margin of error for catastrophe models at anywhere between a factor of 200% and $1,500%. Reinsurers could spend more than $260,000 on such weather models, he said, but common sense is priceless.

GOOD UNDERWRITING

“A good underwriter is better than any model that’s out there,” said Castaldi. “In Canada, you do have significant exposures, but they’re easier to manage because they occur so often.”

A good underwriter will consider climate patterns that have been developing over the past decade or more, Castaldi said. “Ah, remember the Ice Storm of 1998. Now I don’t want to get everybody upset or nervous, but that happened during an El Nino year. Well, El Nino’s back. Let’s hope we don’t have that same situation.”

In 1998, the water temperatures in the east Pacific created two jet streams that merged warm and cold air fronts. The result was a storm that resulted in a three-inch layer of ice that coated eastern Ontario and Quebec.

The 1998 Ice Storm generated more than 700,000 claims – the largest frequency event in the world for the property and casualty industry, Russell noted. It also generated a class-action lawsuit in Quebec that, if successful, could almost double the Cdn$1.2 billion that the storm caused in insured damages. “Of course insurers have a very strong case in that those types of losses were not covered and could be denied insured peril,” Russell said.

If the Ice Storm were to happen again, underwriters would need to factor in that damage losses increase every year based on inflation, demographic shifts and urban development patterns, Castaldi said. “Keep in mind those past losses, because they can occur again. And remember what I said: in every 10 years, [the value of the losses] could double.”

Given the Cdn$1.2-billion in damage losses for the 1998 Ice Storm, underwriters would be looking at paying out Cdn$2.5 billion if a comparable storm occurred in 2008, according to Castaldi’s rule of thumb.

Common sense underwriting is also an antidote to situations the models may be limited in handling. “Things like floating casinos – what were we thinking?” Castaldi said. “We learn that they float, crash into buildings, they sink, and they fall apart.

“That’s common sense underwriting, but how do you model that? Most people put that in as a fire-resistant building on land and they had no losses whatsoever.”

WEATHER PATTERNS

Catastrophe losses have become increasingly weather-related, Castaldi observed. Insured weather losses in Canada between 1985 and 1995 totaled Cdn$1.7 billion; between 1995 and 2005, weather-related losses have totaled Cdn$6.7 billion.

“What should you be concerned about?” Castaldi asked. “If most of your losses are weather-related and hydro logical (flood or drought), the thing that would concern me is global temperatures.”

And those temperatures have lately demonstrated more extremes.”Something is happening,” Castaldi said. “We’re getting warmer.

“In Canada…we’re going to see more extremes. That means more rainfall, more cold, more droughts, so that’s something to be concerned about. Keep it in mind. We might have a good year in 2006, but that could be short-lived.”

Just as climate patterns can help predict coming storms, so the position and shifting of plates under the earth’s surface can tell us about the possibility of a major earthquake happening in Canada.

“There is a 70% possibility that a major earthquake will hit B.C. in the next 200 years,” panelist Mariagiovanna Guatteri told seminar participants.

She said it was entirely possible that a major earthquake of the type that originated in Sumatra in 2004 could also happen in western Canada. The 9.3-magnitude earthquake in Sumatra resulted in a tsunami that killed more than an estimated 300,000 people in Indonesia, Thailand and Tanzania.

Since 1660, Canada has experienced 10 important earthquakes, most of them in B.C., Guatteri said. She said a large event like the Cascadi earthquake that hit the Western Canadian coastline in 1700 happens once in about every 500 years.

In preparing to minimize damage losses caused by a major Canadian earthquake, building construction should vary depending on whether ground movement happens in B.C. or in the Ontario/Quebec region, Guatteri noted.

Low-rise buildings in Quebec are more susceptible to shorter periods of ground shaking; they would therefore be vulnerable to the region’s earthquake patterns, she said. Conversely, high rises would be especially vulnerable to collapse in B.C. because of the longer-frequency shaking in that area, augmented by the province’s soft soil.