Catastrophe modelling experts told actuaries at an industry conference that they need to understand, use and interpret catastrophe modeling results correctly to impact the insurance industry's future loss potential.
Their comments at the Casualty Actuarial Society's 2006 spring meeting in Fajardo, Puerto Rico were reported by CAS.
"Catastrophe models are tools that allow you to evaluate risk within a framework and provide valuable information for effective risk management," said David A. Lalonde, senior vice president, AIR Worldwide Corp.
"However, users need to better understand and test the sensitivities of the models they employ, as well as interpret the results properly," continued Mr. Lalonde.
"Model users must know the strengths and weaknesses of their model and the science and engineering upon which they are based," he said. "They also must utilize the results with care and caution."
While the primary output of catastrophe models is to pinpoint distribution of losses and show the probability of exceeding various size losses, models also provide information on industry- and company-specific portfolios and individual risks, including estimated average annual, occurrence, annual aggregate, deterministic event and historical event losses, CAS noted.
As a result of record losses from Hurricane Katrina, exposure data quality has become a high priority for many insurers, according to the organization's report. Complete and high-quality exposure data are essential to reliably assess catastrophe risk.
Because much of data gathered about risks is collected through older interfaces, such as financial or premium systems, the models may be prone to inaccuracy. "Can insurers afford not to capture this data accurately?" Maria Kovas, director of information services, Toa Re America, asked rhetorically.
Insurers that are unsure about the integrity of their exposure data can check the reasonableness of their model output by benchmarking their catastrophe modeling results against industry losses, advised Mr. Lalonde.
"As a benchmark, a primary insurer that has a 5 percent market share should incur roughly 5 percent of the industry loss after an event," said Mr. Lalonde, according to the CAS report. "Insurers can check this during the modeling process by conducting a market share analysis of their book of business."
© Arc, All Rights Reserved. Request academic re-use from www.copyright.com. All other uses, submit a request to TMSalesOperations@arc-network.com. For more information visit Asset & Logo Licensing.