By MATTHEW BRODSKY, senior editor/Web editor of Risk & Insurance®.
Back in the days of Bill Clinton before e-mail, when children bullied each other in person instead of in virtual chat rooms and adults could escape work by leaving the office, Peter Yanev was working with one of the world's largest insurers on the 8.2-magnitude earthquake in Guam in August 1993.
The earthquake expert and structural engineer gave a presentation to the insurer's staff about damage to some of the facilities on the U.S. territory. He showed the slide of one, a multinational hotel, and mentioned that damage to it probably amounted to at most a 20 percent loss. Keep in mind, Yanev had snapped the hotel's photo and made his loss estimate on a drive-by.
Turns out, this very same property was on the insurance company's books. And one of its adjusters had already visited it and reported back that the hotel was--here is the kicker--a total loss.
After Yanev's little photo and comment, the insurer dispatched its own engineer to the site, and, yes, it was just a 20 percent loss at most.
"We're talking $20 million because I showed one slide and made a comment," Yanev said.
"I got to Chile now, and my sense is that nothing has changed," Yanev added about the monstrous 8.8-magnitude quake of Feb. 27, whose aftereffects he witnessed first-hand.
Now, 17 years since the Guam temblor, two more polarizing presidents and a couple rinse cycles in the hard-soft insurance market swing, and the property/casualty industry still hasn't learned the lesson? Nope, according to Yanev, who co-founded modeling firm Eqecat Inc. in 1994 to help it do so.
"My sense is the insurers do not appreciate the complexity of the issue," he said. "There is a lot of spinning of wheels and a lot of tearing down buildings and a lot of unnecessary losses."
More frightening is the opposite: insurers that are "dead wrong" when they underestimate their potential losses in future quakes, as Yanev suspects they are doing with the Cascadia Subduction zone, the major fault in the Pacific Northwest.
This region is really the only one in the United States where a quake the size of the one that struck Chile in February will happen anytime from tomorrow to a couple centuries from now. And unlike in Chile, which has now-proven, world-class building codes, the building codes in the states of Washington and Oregon and in the Vancouver, British Columbia region are frighteningly inadequate.
Yanev is suggesting that insurers do not quite appreciate just how poorly properties will fare in the Pacific Northwest when the fault snaps.
Now mind you, acquaintances and former colleagues of Yanev will smile and tell you, "Oh, that's Peter." The structural engineer has been on the ground after many, many quakes and has worked with the insurance industry for decades, they admit, but Yanev sometimes perhaps doesn't let the facts get in the way of his passion.
Even your humble author, who covers the industry like a tourist from one guided tour to another, has seen enough to know that today's insurance world is very sophisticated in how it handles catastrophes, with its legions of trained and experienced underwriters and adjusters and its 21st century modeling tools. Perhaps Yanev is a bit too broad-brush, too enthusiastic in his condemnation.
Then again, on this guided tour of earthquake risk and the lessons to be learned from Chile 2010, your humble author has heard from "locals"--experienced insurance insiders--who've voiced comments here and there that could lead one to believe that Yanev isn't out there, that the industry does have lessons to be learned when it comes to earthquake exposure management.
Yes, we've all heard the lesson of building codes: how Chile proved how important and effective they are, how Haiti suffered from their absence. But here are other lessons from the Chile temblor that haven't been properly aired yet.
For starters, there is the fear that insurers are not properly pricing quake risk (right now, of course, one could argue nothing is being priced correctly).
Ivan Gonzalez, director of client markets with Swiss Re, whose potential losses in the Chile quake could total $630 million, was not sure whether competitors in Chile were pricing tsunami with earthquake, for instance.
Launched by the pop of geological plates, local tsunamis rose against coastal towns and ports in Chile near the epicenter, particularly near the town of Constitucion. Gonzalez explained how if tsunami were priced into property insurance in the last two to three years in Chile, those rates would be very uncompetitive compared with markets not factoring in tsunami--and therefore charging less.
A big insurance takeaway from Chile, said Dr. Laurie Johnson, the chief science adviser at Lexington Insurance Co., was "just the significance" of tsunami risk.
The port damage in Chile, said Mohsen Rahnama, who led the development of Risk Management Solutions' earthquake model and headed the firm's team that traveled to Chile days after the event, was more than he'd ever seen.
Insurers might also be too optimistic with their rates in the Pacific Northwest.
Jonathan Hall, executive vice president with FM Global, appears to agree with Yanev on this one, calling his own employer, the Johnston, R.I.-based commercial property insurance specialist, the "lone duck" on Pacific Northwest earthquake.
The words "competitive disadvantage" came up in a conversation with Hall, and it sounds like Hall's employer is charging more for capacity in the Pacific Northwest than perhaps competitors are because of a more cautious approach to the risk.
Ask any property underwriter if they are low-balling their loss estimates for the Pacific Northwest, nevertheless, and they will never admit it.
Not to say Ironshore is one such insurer, but you can imagine many property underwriters responding in the same way as did Tony Mammolite, head of worldwide property at the Bermuda-based property/casualty insurer.
His employer, he said, manages its aggregates in all catastrophe hotspots. It's a matter of simply knowing how much capacity you can afford to write in a certain area and not going north of that number.
"I don't think we're underestimating the danger," Mammolite said in particular about the Pacific Northwest.
Yet, if carriers are so good at managing their aggregate catastrophe exposure hotspot to hotspot, how come so many reinsurers got their pants yanked down in Chile? Reinsurers are supposed to be some of the most sophisticated users of catastrophe modeling software, the tools the insurance industry uses to comprehend how disasters could impact their books of business. Chile, even before February, was a well-known earthquake hotspot. It's the home of the strongest temblor ever recorded, the 9.5-magnitude quake of 1960.
Up until the February event, though, reinsurers were moving capacity into Chile to diversify, said Swiss Re's Gonzalez, all the while not being fully aware of their full exposure.
Remember, Gonzalez' own employer took a greater-than-first-thought nine-digit hit there.
Competitor Munich Re ate an even bigger loss of about $1 billion. Smaller players took big hits, too. PartnerRe's losses topped $300 million, which amounts to 2.5 percent of the Bermuda company's 2009 equity of $7.645 billion. Amlin, the largest Lloyd's underwriter, suffered £123 million in losses, or about 7.7 percent of its £1.593 billion in 2009 total reserves and equity. The Lloyd's market as a whole took a $1.4 billion smack in Chile (out of an estimated $8 billion total industry loss).
Chile wasn't cheap.
"There will be lots of surprises for people," said RMS' Rahnama after he visited Chile.
Reinsurers might not have had a clear understanding of the industrial and commercial buildings they were writing down there, Rahnama also said. They might though, now that the claims process is well under way. Now that it's too late.
This lack of clear understanding could be because, thankfully, big quakes don't hit people- and property- packed areas all that often.
"It's been a number of years since the industry has been tested," said Sanjay Godhwani, executive vice president and property division executive at Lexington.
Another reason: Many properties suffered nonmodeled losses, or losses reinsurers didn't see when they were first pricing business in Chile. Heavy losses resulted from nonstructural components, failing ceiling tiles, displaced equipment, and ruptured heating and cooling pipes.
"What we're seeing is that there were tremendous nonstructural impacts in otherwise well performing buildings, like hospitals," said Johnson, who also visited the quake damage soon after the event. Hospitals had to be evacuated because nonstructural walls were compromised, for instance, and ceiling tiles had dropped and led to water damage, she said.
"And they sort of fall in this gray zone," she added, explaining that when a risk manager is evaluating the structure and contents of a facility, these sort of nonstructural elements often get left out.
Embracing the impact of nonstructural elements has become "one of the most important aspects of earthquake engineering," said Andy Thompson, head of the Americas risk consulting business at global design and engineering firm Arup.
Yet in Chile, Thompson, a structural engineer, found facilities that performed "absolutely fine" on the outside but experienced nonstructural damage on the inside that could lead to weeks or months of business interruption.
Business interruption did not stop there for commercial property owners in Chile and their insurers and reinsurers. Business interruption can have a "cascading effect," said Kate Stillwell, earthquake product manager for modeling firm Eqecat Inc., an effect that isn't easily captured in the mathematics of her firm's models.
That inadequacy can have its own cascade. It means reinsurers using those models might not have understood this aspect of their exposure either, which means they didn't properly price the risk.
Other large business interruption domino effects in Chile resulted from damage to infrastructure. Guillermo Franco, the principal engineer at rival modeler AIR Worldwide who led the firm's damage survey into Chile after the quake, said that bridges constructed under the current building code performed well; those that were not did not. Even some sections of the Pan-American Highway collapsed. But equally crippling was the loss of electrical, telecommunications and water supply lines.
"A lot of business interruption losses will be had because of the failure of these systems," he said, adding in a forthright way that commercial catastrophe models do not capture business interruption exposure caused by these lifeline losses.
Let's drive the point home so we can understand the implications here. The state of Washington started a major transportation retrofit, according to Lexington's Johnson, but the state still has double-decker highways like those that failed in California in 1989 during the Loma Prieta earthquake.
"The transportation outage would really be devastating to the economy long term," she said.
Even earthquake insurers operating in California could suffer the consequences of not understanding the cascading effect on insured losses of infrastructure hits, and California is the epicenter of U.S. quake risk. If there were one quake hotspot insurers should be prepared for, it'd be this one. You'd think.
Johnson worked on a 8.0-magnitude quake simulation for Southern California in 2008 as part of the Great California Shakeout. The direct economic damage was projected to be in the $100 billion range, much of it based on the loss of key infrastructure pieces.
All of the major water supply infrastructure for the south crosses the fault zone, and major port sites are there, too.
Johnson wonders what would happen if the projected scenario became reality, and Southern California went without water for months on end, or if ports shut down and oil couldn't be brought in.
Sure, many business interruption losses might not be covered in a commercial property policy, unless perhaps ingress-egress were included and triggered, say, or perhaps the insured bought contingent business interruption coverage.
In California, many residents and businesses have no quake insurance. Last year, residential take-up rates, for example, were estimated at only 12 percent.
In many cases, the commercial and governmental interests in the affected region could be swallowing the resulting losses. Not insurers.
True, it's up to risk managers and other leaders at these companies to design business continuity plans that take infrastructure failure into consideration, said Charles Bauroth, technical operations manager for Liberty Mutual Property.
In Chile, some businesses weren't able to carry out business continuity plans because they didn't take into account infrastructure, said John Merkovsky, global practice leader for Marsh Risk Consulting.
A separate article could be written about lessons for corporate risk management out of Chile (and will be, to be posted on riskandinsurance.com beside this article).
But the point here to ponder is about lessons learned by underwriters, about uncertainty and about how insurers do not experience earthquake losses enough to understand their full exposure.
September 1, 2010
Copyright 2010© LRP Publications