Few operational aspects of the insurance industry rely on cutting-edge technology to the extent that catastrophe modeling does.

Property/casualty insurers, reinsurers and reinsurance intermediaries all employ an array of sophisticated models and tools to get a better sense of the risk presented by earthquakes, floods and windstorms to lives and property.

The insurance industry is greatly dependent in these endeavors by well-regarded independent catastrophe modeling firms such as Boston-based AIR Worldwide Corp., Newark, Calif.-based Risk Management Solutions Inc. and Oakland, Calif.-based EQECAT Inc.

The firms, in turn, continue to improve their offerings.

EQECAT folded a litany of improvements into Version 13 of its Risk Quantification & Engineering platform, including an enhanced financial model, an improved user interface and a uniform database schema in April. The new version also included updated earthquake vulnerability models for countries including Australia, China, New Zealand, Portugal and Spain, and updated tropical cyclone models for countries including Australia, India, Mexico and Venezuela.

Speaking at an event held in Chicago in June to introduce users to changes in Version 13, Tom Larsen, senior vp and product architect for EQECAT, said the new version is the result of three years of work by the firm and incorporates the latest in scientific research as well as new loss data. For example, he cited new insights derived from correlating soil composition to corresponding losses to insured structures for earthquake-ravaged Christchurch, New Zealand. “Soils are key to understanding potential damages,” he said.

Given the breadth of changes included in the model, Mr. Larsen noted that the company has a dedicated change management model to assist its clients in adopting the new model.

Model changes affect underwriting

Indeed, changes to catastrophe models can cause significant ripples in an insurance company's operational posture by necessitating adjustments to underwriting practices, capital requirement and reinsurance protections.

This was evident last year, when RMS released Version 11 of its U.S. windstorm model. After considerable complaining by its clients, the company subsequently issued a press release explaining that the increased volatility in models was a direct result of improvements in underlying calibration data and modeling methodologies.

“RMS gains no advantage in over- or understating the risk and will not consider selectively withholding modeling advances to dilute or phase changes in results,” the release said. “While model change can bring significant challenges, we believe companies should have access to all new and relevant information known to and reviewed by us so they can adjust underwriting decisions and capital requirements appropriately. The alternative would lead to compromised decisions and surprises when catastrophes strike, and has associated liability implications.”

It is this inherent volatility in models has led many underwriters to develop a more nuanced view of risk by blending outputs from multiple models, said John DeMartini, New York-based leader of the catastrophe risk management practice at Towers Watson & Co.

“All these incidents from 2011 and the change in the RMS model got people thinking that they need to develop their view of cat risk,” he said. “We should not be entirely model-dependent; we have other tools and analyses available to us.”