Conference Raises Questions For P-C Insurers

NEW YORK–Property-casualty insurers looking for clear-cut guidance for their catastrophe risk decisions were given much to think about after yesterday's session sponsored by The Bond Market Association.

Two meteorologist's, Kerry A. Emanuel, professor of atmospheric science at the Massachusetts Institute of Technology, and Stanley Goldberg, a meteorologist with the hurricane research division in Miami for the National Oceanic & Atmospheric Administration (NOAA), presented contrasting views on the future of hurricanes.

Mr. Emanuel said increased hurricane activity witnessed in the last few years is a byproduct of global warming caused by the release of man-made carbon emissions. In his interpretation of the data, duration and intensity of hurricanes is increasing primarily because of the warming of the sea surface.

The increased intensity of hurricanes corresponds with the increased amount of carbon emissions in the atmosphere going back to the 1970's, he said. Taking a longer view of storms, going back to the 1870's, he dismissed the notion of cyclical patterns to hurricanes, saying there are variations in hurricane activity, but not cycles.

“Any long-term risk assessment must account for climate change and carbon gases,” he concluded, adding that this is the consensus view of the science community.

Mr. Goldberg countered Mr. Emanuel's views, noting the two have discussed their differing opinions on occasion. He disputed that a consensus in the scientific community finds that “there is more to it than greenhouse gas warming.”

The NOAA scientist made the point that the primary driver of hurricanes is not ocean warming but “vertical sheer”–upper and lower level winds–which can tear the top off of a tropical depression and prevent its formation. In his interpretation of the data, ranging from 1870's to 1998, there is clear evidence of cyclic activity, though there is little evidence of atmospheric change over the early period.

He argued that recent observations of growing intensity of storms may not be an actual increase, but instead may reflect better abilities to collect precise data. He added that only recently have scientists been able to collect such precise, active storm year data. Mr. Goldberg added that more “re-analysis of data is required” before conclusions can be reached.

No matter which opinion is correct, said Mr. Goldberg, there is an urgent need for hurricane preparedness.

“Hurricanes happen,” Mr. Goldberg warned. “We need to be focused on preparation no matter who is right or wrong.”

For catastrophe modelers providing information for insurers to make their risk assessments, new scientific data and claims experience has lead to changes in some forecasts.

Dennis E. Kuzak, senior vice president with EQECAT Inc., S. Ming Lee, executive vice president with AIR Worldwide Corporation, and Peter Nakada, senior vice president, RMS Consulting, discussed ways new data has altered their assessments and increased loss and frequency estimates.

Mr. Kuzak pointed out that frequent disputes over conclusions of data among scientists can result in “violent” disagreements.

“We are inundated with scientific information,” he said of modelers' work. “The ground rules are that the models need to be constrained and consistent with the records and science.”

The modelers' loss estimates for a variety of catastrophe events could differ by billions of dollars, Mr. Kuzak said. For example, one estimate from RMS, looking at long-term cat risks as much as 50 years out, speculated the industry would require $120 billion in capital and need to add $82 billion to its book.

Differences of opinion, explained Mr. Nakada, are due to a “healthiness” in interpretation that reflects different data and the complexity of application.

From a rating perspective, Gary Martucci, director, financial services group, Standards & Poor's, said when using these estimates the service tends to use the most conservative analysis in its ratings. He credited modelers with using the best science available, noting that there is no favoritism in approach.

Stephen Lowe, managing director with Tillinghast Towers Perrin, was critical of those who use the modelers and fail to recognize their limitations.

“It is important to recognize that models are estimates,” Mr. Lowe said. “There is a lot of opportunity for error, and a lot of them incorporate uncertainty. The insurance industry needs to upgrade its use of models and be more sophisticated in its use of them.”

He said any failure placed on the modelers is actually the failure of the users to incorporate the models properly into their own risk analysis.

“It's not the fault of the modelers, but the issue of their use, and [the modelers] have borne unfair criticism,” he added.