The guilty verdict against six scientists and a government civil servant for manslaughter in an Italian court may be a foreshock that portends a bigger seismic event ready to strike the catastrophe modeling industry.
Holding modeling firms financially responsible for events that fall outside modeled loss estimates is moving out of boardroom chat into real discussions among reinsurers and other model users that have faced frustrating catastrophe losses over the past several years.
“It really shocked me,” says Katherine Stillwell, manager of earthquake products for EQECAT said when describing being questioned about modeler “liability” during a industry event two weeks ago. “It’s the first time I have ever heard the question asked in a public forum. I think it will be a point of discussion in the future, and possible heated discussion.”
Questions surrounding scientific — and by extension — catastrophe model liability moved to the forefront of industry conciseness when the “L’Aquila verdict” was reached last month. Three scientists and government officials were found guilty of manslaughter, fined $10.2 million and could face six years in prison for what the Italian judge argued were falsely reassuring statements in advance of the 2009 earthquake that killed 309 people.
Photo Alessandro Giangiulio via Flickr used under a Creative Commons Attribution-ShareAlike license
L’Aquila was hit my a 5.8 magnitude earthquake after the scientists told a city risk-assessment commission that they were unable to make a detailed prediction about whether ongoing several foreshocks would portended a larger event. Just prior to the major shock a civil servant was quoted by local media as playing down the threat.
Stillwell says that while the L’Aquila verdict may be the focus of sensational headlines that may be ultimately overturned in Italian courts, it is a symptom of a wider backlash against the role of science and models focused on disaster risk.
“We are seeing more questions from clients that focus on the difference between [model] output and experience,” Stillwell says. “That leads to questions regarding liability for those losses.”
Despite the questions, third-party catastrophe modeling firms work in a sophisticated market where insurance and reinsurance firms should understand the role of “probabilistic models,” argues Dr. Robert Muir Wood, chief research officer for catastrophe modeling firm Risk Management Solutions (RMS) in London.
“We work in a business to business environment by providing information to the business market,” Muir Wood says. “We are always very clear that we are developing probabilistic information and we always refer to the data in a probabilistic perspective. We have pages and pages of legal wording to that affect.”
In fact, Muir Wood points out, the questions of scientific liability and catastrophe risk was addressed in the U.S. courts several decade ago after Hurricane Audrey struck in 1957. The hurricane hit Southwest Louisiana and Southeastern Texas with what new reports called “massive storm surge.”
At least 500 people died following the storm, and over 100 families filed a lawsuit against the National Weather Bureau (a precursor to the National Weather Service) arguing that “inadequate warnings.” After several cases were rejected on appeal, all of the lawsuits were eventually withdrawn, according to the book Hurricane Audrey: The Deadly Storm of 1957.
That is why Muir Wood says he has not been approached by clients of potential liability and that it has not yet hit his radar.“Many times were are asked to deliver certainty in an uncertain event,” Muir Woods says. “But In the context where we provide advice, the situation is very different.”
Stillwell, however, sees a more problematic future for the catastrophe modeling industry and that the discussion of liability may go “one of two ways.”
“My hope is the general public will recognize the difference between the known and unknown,” Stillwell says. “But the more pragmatic side is that the issue of liability will grow and modeling companies will become more cautions about the volume and type of information that is sent out and that will stifle pubic discussion of natural hazards.”
