He noted that the 2006 season stabilized the catastrophe market. For example, he said some insurers in 2006 increased windstorm deductibles to 5 percent–up from 1-to-3 percent in prior years–and imposed higher minimums.

Insurers are easing deductibles in some specific geographic areas, like Houston or surrounding Harris County, Texas, he said. However, buyers with heavy cat exposures–particularly California earthquake–are seeing increases with renewals.

In general, Mr. Ellis said that renewals with moderate cat exposure are seeing flat to slight decreases. On the other side are buyers with “little or no cat exposures who are witnessing a competitive insurance market.” Much like the rest of the p-c industry, these buyers “are seeing decreases of 5- to 15-20 percent on average,” Mr. Ellis said.

He explained that although there is “no real difference in how cat and non-cat-exposed companies should approach the renewal process,” to obtain the best rate and renewal it is essential that buyers:

o Have complete underwriting information, including values, loss history, construction and detailed address.

o Provide information to underwriters in a risk management solutions (RMS) downloadable format.

o Begin preparing renewals as early as possible to investigate all options.

o Use the worldwide insurance marketplace to access capacity–not just the United States.

He said that insurers are competing for non-cat property business as a way to assure premium growth and spread of risk, noting that the non-cat marketplace is the “most flexible it has been in years.”

Mr. Ellis added that international property markets “are seeing an even softer non-cat market than we are seeing in the U.S., providing potential cost and coverage benefit for those companies that engage in these markets.”

The Japanese and Singapore markets, for example, are “eager to write U.S. businesses today,” he said.

With the Terrorism Risk Insurance Act (TRIA) scheduled to expire at the end of this year, the terrorism insurance market is experiencing uncertainty, he observed, noting that despite a changing and uncertain marketplace, “terrorism take-up rates continue to climb to record levels.” Mr. Ellis said that nearly six out of 10 companies “have decided that property terrorism insurance is worth purchasing.”

In the face of significant property price increases, he said that more companies upped their purchase of terrorism coverage in 2006 than in 2007. While TRIA's looming expiration has caused some uncertainty, this has not affected the property terrorism market to the same extent as was the case in 2005, when the original act expired, he said.

“There is tremendous optimism that an agreement will be reached, and TRIA will be extended in some form, based largely on favorable comments by Congressional leaders,” he said.

Unlike in 2005, he noted, “we are not seeing as many insurers insisting on sunset clauses, although we expect that as the year progresses, some markets may add these endorsements more often.”

Overall, he said the property-casualty market place showed a net gain on underwriting of $31.2 billion in 2006, compared with a net loss of $5.6 billion in 2005, according to the Insurance Services Office, (ISO).

Mr. Ellis said there was a huge improvement in the p-c industry's combined ratio for 2006, which stood at 92.4, compared with 100.9 in 2005. The ratio is a measure of losses and other underwriting expenses per dollar of premium paid.

In a discussion about natural hazards modeling, Fletcher A. MacGregor, managing director, property loss control in Detroit, said, “I've seen buyers reduce their premiums in the neighborhood of $500,000 or so.” While this amount won't happen in all places, he said, “the savings can be very significant,” and buyers should pay attention.

He noted that modeling capabilities are not only limited to hurricanes, but “any kind of windstorm, earthquake, terrorism, even workers' comp can be modeled.”

Mr. Macgregor said that rigorous standards go into modeling. For hurricanes, for example, “these companies went back as far as 1900s. What they developed was a correlation between the storm characteristics and subsequent damage resulting from the event.”

Using this data, he said, hypothetical storm sets were developed to fill in coastal areas with no specific history.

“This information is the foundation for the models,” he said. “Once you have that, a property or a portfolio of properties can be subjected to the simulation of thousands of storms and storm tracks and the incremental damage.”

Loss levels can be determined from this information, he concluded.