Hurricane Katrina could cost more than $100 billion, according to the latest estimate from Risk Management Solutions. On Tuesday, risk-modeling firms placed the potential costs at up to $25 billion. Where do these numbers come from, and why do they vary so much?
Teams of mathematicians, statisticians, meteorologists, and structural engineers compile the figures. Risk-modeling companies predict damage by comparing weather forecasts with detailed information about what lies in a storm's path. Using data from the National Hurricane Center, they estimate the wind speeds generated by the storm at various locations. The damage analysis typically addresses entire neighborhoods at a time; for each ZIP code, the modelers know the total value of buildings in the area, what percentage of them are residential, how many of those are wood-frame houses, and so on.
Once the modelers make a guess as to which buildings will be affected, and by what wind intensity, they can start to project the damage—and assign dollar amounts. For this they use "vulnerability functions," which derive from historical data about previous hurricanes. If a residential neighborhood in New Orleans were likely to be hit by Katrina's 100-mph winds, the damage estimates might come from the insurance claims made after 100-mph winds hit a similar neighborhood during Hurricane Charley or Hurricane Andrew.
Historical comparisons don't work for unique or unusual structures in harm's way. When the insurance companies saw Hurricane Floyd approaching Florida in 1999, the potential damage to the space shuttle couldn't be determined from past experience. Since shuttle damage could contribute a significant amount to the total cost of the storm, the modelers assessed the risk informally.
In general, information about previous storms comprises claims for all sorts of losses, like structural damage, destroyed equipment, and even interrupted business. The standard vulnerability functions incorporate all these factors into the final estimate. Overall projections can refer either to the insured damages or to the total costs of the storm. A $9 billion estimate reported in the news on Tuesday referred to insured and uninsured damages from the wind alone; the new $100 billion estimate applies to the total cost including flood damage. These numbers reflect damage to the local economy and infrastructure, but they don't account for global effects—like the dramatic increase in gas prices.
What good are these numbers? Insurance companies hire risk-modeling companies to help them prepare for the aftermath. They want to know how much money they'll need to settle claims, and how many agents they should send into the field. They don't care so much about the overall numbers—for the most part, those are computed for the risk-modeling company's press releases. Instead, the insurance companies use the same data and vulnerability functions to create estimates about their own specific liabilities.
Explainer thanks Rick Clinton of EQECAT and Shannon McKay of Risk Management Solutions.