Market speculation that a current upturn in property rates will be short-lived fails to take into account the likelihood that intense hurricanes will continue for the long term, a top insurer said last week.

And catastrophe models used by insurers and reinsurers need to be ?recalibrated? to take into account the change in storm activity, said Aspen Insurance Holdings chief executive officer Chris O?Kane.

Mr. O?Kane said his company does not expect the impact of the severe storms that ravaged the Gulf Coast and Florida this year to be short-term.

?If our analysis is correct then we are facing a long-term if not permanent change in catastrophe loss expectancies and this will require a permanent solution. Putting it bluntly, prices will go up and they will need to stay up for a very long time,? he said.

Aspen last week reported a third quarter 2005 loss of $362 million or $5.22 per share.

Mr. O?Kane told analysts during the company?s third quarter earnings call on Friday that in a year of exceptional unprecedented storm activity in the Atlantic following four land falling hurricanes last year, the fundamental challenges for the industry are only gradually emerging.

When the full impact of Katrina emerges in the beginning of 2006, Aspen expects a meaningful reduction in capacity, dramatic increases in rates and major changes to terms and conditions. Already there is evidence of reduced capacity with some carriers offering significantly smaller lines.

Mr. O?Kane also noted that the 2004 and 2005 losses put into question the fundamental assumptions about the total amount of insurable value exposed to catastrophe risk especially in light of an increasing number of people and buildings on America?s coast. Questions also exist about the the frequency and severity of hurricane activity and the reliability of catastrophe pricing and accumulation models currently in use.

?There is a widespread view that the so-called catastrophe models are defective and need to be overhauled, replaced, or even abandoned altogether. In our view some of these concerns are overstated. One of the biggest problems we see in the marketplace is unsophisticated use of cat models,? Mr. O?Kane said adding that the marketplace sometimes over relies on ?black box techniques to solve complex issues which model designers have not addressed or offer only very imperfect solutions for?.

?Within Aspen we constantly remind underwriters to price for non-modelled perils such as winter weather or inadequately modelled perils such as brush fire or fire following earthquake,? he said.

Another major problem in the use of models stems from inputting out-of-date or inaccurate exposure data.

?You are not going to price next year?s exposure correctly if you just use last year?s data,? he said.

Aspen moved this summer to address ?significant defects? in approaches to catastrophe modelling by commissioning research to understand the apparent upturn in hurricane frequency and severity.

?We believe that we have now succeeded in recalibrating our model to reflect this change while standard models continue to base annual loss expectancy assumptions on the long term average.

?We have also addressed a surprisingly complex phenomenon known as demand surge and we have now established a better correlation between the degree of demand surge and the size loss,? he said.

Working with external scientific consultants, the company is expecting 20 percent more frequency in the Gulf of Mexico, 25 percent more frequency for Florida and 43 percent more frequency for the East Coast. It is currently expecting an average 7.5 percent more severity.