The RMS model called Risklink 4.3 U.S. Hurricane and U.S. Western Region was introduced in February of this year.

“There are some areas where results have gone down and some where theyve gone up,” said Robert Muir-Wood, head of global risk modeling at RMS, which has offices in London and Newark, Calif., explaining that the change in risk perspective depends on region and return period of loss.

William Ricker, president of RenaissanceRe Holdings in Bermuda, said, “Its actually caused a fair amount of upheaval” because it is changing the perception of what a ceding companys exposure is.”

“Most of the big companies are buying additional top layers or have already done it on the first of April or the first of July,” said Russell Merrett, treaty underwriter for Syndicate 33, which is managed by Hiscox plc, a Lloyds managing agency.

“I found myself being rather busier than I thought at this time of year because the RMS 4.3 model has rattled everybody a bit in Northeastern exposures. So theres quite a bit of extra cover being bought,” said a broker who didnt want to be named.

Overall the frequency is reduced but the severity is increased, according to Chris Smelt, underwriting partner with 2791, which is managed by MAP, a Lloyds managing agency.

Although its a very uncertain science, “what actually happens in modeling is that people get a certain degree of false comfort once they see the numbers for a few years,” said Mr. Ricker. With the new model from the same vendor, ceding companies are getting numbers that are 100 percent higher or more, “and now theyre in a bit of a tizzy on what to do.”

Mr. Ricker said he thinks the changes in RMS earthquake model are actually more significant in changing behavior than the hurricane model at this point, depending on a ceding companys portfolio. An earthquake writer that has been using the RMS 4.2 model to select which risks to write and which to avoid for the last five years may find that when it runs its portfolio through the 4.3 model, much larger numbers appear, he said.

“Youd have to go buy reinsurance or significantly change your underwriting practices,” he said.

As for the windstorm model, Mr. Ricker said, RMS in the past, compared to the other models, “had smaller losses in the tail and more frequent losses in their overall catalogue,” he said. (The tail of the model relates to severe losses, which have less frequent return times.)

“However, their new version appears to have larger losses in the less frequent return times, or the 1-in-100 or 1-in-250-year losses, for example.”

Mr. Ricker emphasized, however, that the average annual loss, or the long-term loss expectancy to portfolios, hasnt changed that much,” he said.

“If your less frequent events are larger and your annual expected is about the same, by definition your more frequent events have to be a little bit smaller or your frequency has to be a little bit less,” he said. “But that actually makes them more consistent with the other modelers.”

Mr. Smelt said models are a guide, but not an exact science, and all modeling companies have slightly different results. He noted that MAP has developed its own in-house model based on market losses, which is a more prudent way to look at probable maximum losses.

As a result of this in-house modeling, Mr. Smelt said MAP is of the view that a lot of cedents should be buying almost twice as much reinsurance as they buy today. “We are reinsurers and were probably a little more conservative than most, but theyre definitely not buying up to their PMLs at the moment.”

Given the inexact nature of the science, Mr. Smelt said, one has to question why people are relying so heavily on RMS when the numbers can change so easily.

“I think its quite hard on the cedents because reinsurers will say thats the number because its out there in the public domain. Its said with great certainty,” said John Hamblin, active underwriter with Syndicate 2010, which is managed by Cathedral Underwriting, the Lloyds managing agency.

“Your 1-in-250-year loss is now Y. Well yesterday it was X and the exposures havent changed one dollar. I think its very difficult,” Mr. Hamblin said. “Its also difficult on ceding companies because the rating agencies buy into the whole thing,” he said. The simple truth is you cant buy enough reinsurance, he added.

“The models just continue to get beter and better. Its an evolutionary development,” said Mr. Merrett. “In five years, the models will be better and will change again, which might mean higher or lower numbers.”

A lot of people are very upset with RMS and cant believe they have to take this change seriously, he said. “On what grounds? Do they think RMS has a vested interest in making the numbers higher? I believe RMS may be right or may be wrong, but theyre making a sincere effort to estimate the impact of losses over time.”

Mr. Merrett thought that some of the rationale behind the extra purchasing of reinsurance also has to do with the fact that “some of the big companies have really enjoyed growth in their exposures over the last 18 months to two years.”

He explained theyve been writing more business because of the exit of competitors, such as Reliance and Kemper. “I dont think its just the model change.”

Reproduced from National Underwriter Property & Casualty/Risk & Benefits Management Edition, August 25, 2003. Copyright 2003 by The National Underwriter Company in the serial publication. All rights reserved.Copyright in this article as an independent work may be held by the author.