Insurance Method For Hurricane Rates Draws Skepticism
Dec. 19--Just weeks after monster storms Katrina and Rita devastated the Gulf Coast, a risk management firm flew a handful of the world's top hurricane scientists to Bermuda and its famous pink sandy beaches.
There, the scientists were asked to predict Atlantic hurricane activity for 2006 to 2010.
They estimated hurricane activity would be significantly higher than the long-term, historical average, calling for 40 percent more landfalls of major U.S. hurricanes than normal during the period.
Shortly thereafter, the firm, Risk Management Solutions, or RMS, increased its "modeled insurance losses," which help establish rates for insurers and re-insurers, by 40 percent.
Outside insurance circles, this expert elicitation process isn't well known. But it has at least indirectly affected homeowner insurance rates for many coastal customers.
Now, in addition to some insurance regulators, scientists who have participated in the process have begun questioning the method.
Some participants in the latest meeting, held in October in Miami, say there's not enough evidence available to produce a skillful forecast of five-year hurricane activity. And at least one is worried that the exercise might be used to justify increased insurance rates.
"For all I know, they believe that they can predict activity five years into the future, but the science does not support this belief," said Roger Pielke Jr., a professor of environmental studies at the University of Colorado who participated in this year's meeting. "How can this be anything other than putting a scientific imprimatur on a question that science has yet to be able to answer?"
Whether the elicitation process has the power to influence the lowering of rates, in addition to raising them, will soon be tested.
Prediction revised
During the October gathering, the scientists lowered expectations. Their revised estimate called for just 20 percent more landfalls of major hurricanes -- defined as a Category 3, 4 or 5 -- during the next five years than the long-term average.
Historically, about two major hurricanes strike the United States every three years.
So far, RMS hasn't acted to lower its modeled losses based upon the most recent elicitation. "It is too early to conclude what impact" the new estimate will have on the 2009 models, the company said.
Several firms provide models to insurers to help them understand their financial risk from hurricanes. Traditionally, these models have estimated risk based upon the long-term average of hurricane activity.
But after the disastrous 2004 and 2005 hurricane seasons, RMS and others sought to develop what they termed "medium-term," or five-year models that reflected the greater-than-normal activity since 1995.
Because there was no consensus on how best to model hurricane activity over five years, RMS turned to scientists, said Christine Ziehmann, director of model management at RMS. "The expert elicitation approach avoids the danger of picking just one forecast model by building multiple models using historical data as well as predictions of future conditions, and gathering external expert judgment on which models are more suitable than others," she said.
By early 2006, three risk-modeling companies -- RMS, AIR Worldwide and EQECAT -- had released ""medium-term" models. Some of these were used by reinsurers -- companies that provide insurance to insurance companies -- to raise rates.
However, companies such as Allstate and State Farm, which sell homeowner policies, must go through state regulatory processes to receive higher rates.
In Texas, the new models could be used as part of an insurance company's justification for higher rates in applications to the Texas Department of Insurance. Although the models are just a piece of the puzzle, it's notable that homeowner rates for coastal parts of Texas have risen by about 6 percent since 2005's Hurricane Rita. The statewide average has fallen by a similar amount.
In Florida, a state more exposed to hurricanes than Texas, a special commission must approve the models that insurance companies use to set rates.
The Florida commission chose not to accept RMS' five-year model. Had the commission allowed insurers in Florida to use the model, homeowners' rates would have risen by as much as 40 percent.
"RMS has been unhappy, to say the best of it, with the commission's standards," said a member of the panel, Howard Eagelfeld, who is also an actuary with the Florida Office of Insurance Regulation.
Some participants in RMS' most recent elicitation appeared to share some of the same apprehension about what they were being asked to do.
At the Miami meeting, the hurricane scientists were told to rank 20 different models. One model, for example, would project future activity based on historical storm counts. Another assumed warmer seas would produce stronger hurricanes. RMS ultimately used a combination of the best models chosen by the experts to develop its medium-term estimates.
Losses overestimated
"I understand they need an answer on a particular number of storms," said Gabriel Vecchi, a research scientist at the government's Geophysical Fluid Dynamics Laboratory who participated. "But this isn't at the cutting edge of what the scientific community is trying to do. It's something that's really at the cutting edge of what people are trying to understand how to do."
The medium-term models developed by RMS and others received further criticism this month when an independent risk analysis firm, Karen Clark & Co., found that the models significantly overestimated hurricane losses for the 2006 through 2008 seasons.
The medium-term models developed by the three firms predicted insured losses of $37.2 billion to $42 billion for those years. The actual cumulative losses during that time were $13.3 billion.
Nonetheless, RMS has confidence in its approach and, despite the results of the recent expert elicitation, has not signaled a significant change in its medium-term models.
A.M. Best Affirms Ratings of Blue Cross and Blue Shield of South Carolina; Assigns Ratings to InStil Health Insurance Company
J.G. Wentworth Completes $74.6 Million Securitization of Structured Settlement-Backed Notes
Advisor News
Annuity News
Health/Employee Benefits News
Life Insurance News