Like many storms, Hurricane Dennis loomed larger in hype than in reality. But Dennis' landfall on July 10 as a Category 3 hurricane had implications for hurricane modeling and catastrophe management in general. Its unprecedented early arrival as a major storm will affect how storms are predicted and how insurers prepare for them. Dennis also provoked an early debut of technologically driven tools and strategies demonstrating the evolution of CAT-related solutions, or simply representing such technology's increasing availability.
Further, Hurricane Dennis has served to corroborate predictions of an unusually severe hurricane season. It also provided an opportunity to put in practice what was learned from last year's Atlantic hurricane season.
One of the key lessons of the 2004 season was the need to deepen information and services on the uncertainties that exist in real-time hurricane loss estimation, according to Kyle Beatty, a meteorologist with catastrophe risk management solutions firm RMS (Newark, Calif.). "Every hurricane is unique, and observations of a storm's hazard characteristics are sparse, requiring assessments outside the range of conventional data sources to narrow model uncertainty," he says.
Beatty adds that RMS is seeking to narrow the uncertainty factor through more-efficient and more-detailed tools for 2005 that can infer the intensity of a hurricane in places where wind observations are unavailable. Also new with Dennis was RMS' deployment of high-resolution aerial video imagery from Long Beach, Calif.-based ImageCat. The information it provides was used to validate the storm's extent and intensity in areas with a significant amount of insured property.
RMS competitor AIR Worldwide (Boston) this season is introducing improvements in the way it estimates insurance losses by quantifying the probabilities related to each loss estimate, according to Jay Guin, the vendor's vice president of research and modeling. "It's one thing to give out a range of losses, but if one can associate probabilities with it, that becomes much more useful information," he says.
If, hypothetically, AIR predicts a range of losses for a given storm to be between $1 billion to $5 billion, it now can say, for example, that a 50 percent chance exists that the losses will exceed $3 billion, but only a 10 percent chance that the loss will exceed $5 billion, Guin explains. This refinement in estimation can further assist insurers trying to allocate both claims management and financial resources.
Ponte Vedra Beach, Fla.-based Sunshine State Insurance ($81.4 million in 2004 premium) uses information gathered from AIR's ALERT online, real-time loss estimate service to aid it in allocating its catastrophe management resources, according to Michael Cratem, vice president, Sunshine State. "We put that information quickly into the hands of our vice president of claims, who then can see which policies and which geographic areas are affected by the storm," he says. "From managing the finances of the company right down to getting people out into the field, it helps us to be ready."
Sunshine State complements information from ALERT with data gathered on the ground. "If ALERT says we can anticipate somewhere between 300 and 500 claims and the severity is 'X,' we set our reserves accordingly," Cratem relates. "If our adjusters report seeing damage not quite as bad or worse, then we make other changes." However, those adjustments tend to be minor, Cratem insists. "For the most part, the losses tend to be in an acceptable range of what the [AIR] model says - it's amazing how closely we manage to that number."
Still, AIR continually is looking to improve the accuracy of its models. Along with other AIR clients, Sunshine State collaborates with the vendor by providing actual loss information, which AIR then uses to evaluate its predictions and increase its store of historical data for future predictive purposes.
Anthony O'Donnell has covered technology in the insurance industry since 2000, when he joined the editorial staff of Insurance & Technology. As an editor and reporter for I&T and the InformationWeek Financial Services of TechWeb he has written on all areas of information ... View Full Bio