Fitch Re-evaluating CAT Methodology

Fitch Ratings is re-evaluating insurers' use of catastrophic risk models, saying the current programs are inadequate for developing a true picture of a carrier's capital requirements.

Peter Patrino, a member of Fitch's insurance group based in Chicago, said: "The old methods for evaluating catastrophe risk are just that, they have gotten a bit old. The new process and thoughts that Fitch will bring to the evaluation of CAT risk will look at the entire tail of risk distribution and integrate that risk with the insurers' overall capital needs."

The company's comments came today during a telephone press conference to discuss Fitch's special report, "New Thinking on Catastrophic Risk and Capital Requirements."

Fitch said recent events caused it to make the re-evaluation after insurers racked up losses of $70 billion over the past two storm seasons, noting that result "is certainly worthy of our attention."

Mr. Patrino said the report touches on why CAT analysis is such difficult prediction work. The events are unusual and difficult to predict loss for, he said, but there are other reasons that make forecasting difficult.

Mr. Patrino said there is poor public disclosure of catastrophe profiles by insurers and shortfalls in regulatory catastrophe models. The report also finds inconsistent terminology related to catastrophe risk, leading to different evaluations of capital among insurers.

The reliance on catastrophe modelers also comes into question in the report, he said.

"We believe that the CAT models are terrific tools," he said, "but other fundamental actions are required by insurers, including poor underwriting oversight and prudent reinsurance utilization.

"When [carriers] stand alone, there is risk that insurers are relying too heavily on CAT models and the outcome they produce. We believe that insurers that supplement underwriting with CAT models are much more effective in their management of catastrophe risk."

He said Fitch is skeptical of the data put into models and the ability of insurers to use the models in their underwriting.

"Any sign of improvement in this area we would think would be a good development, but we are not too sure we will be coming to that conclusion anytime in the near future," Mr. Patrino observed, adding that the ability to use the models varies from insurer to insurer.

How Fitch will change its analysis of catastrophe will be to make a long-term examination of catastrophic risk. Instead of looking at one incident along a time line, such as a 250-year storm, it will take a point of loss and form a time line to the worse case of loss. The analysis will be similar to what is currently done with annuities analysis, he said.

The capital risk will be simulated with statute risk and reserving risk, to come up with a profile. The risk profile, that will translate into a rating, will be different among insurers, he said.

Fitch will use a product from AIR Worldwide in Boston to perform this analysis. Incorporating other data as needed, Mr. Patrino said Fitch is confident it can produce better ratings scenarios. This updated evaluation should be completed within six months.

"Our model is only a model and is not the exclusive answer on CAT risk as it relates to Fitch's insurance ratings," said Mr. Patrino. "We will supplement this approach with other approaches we have used historically in coming to our overall ratings conclusions."

Additional information and a copy of the report are available at www.fitchratings.com.