Cat Modeling, Forecasting Tools More Sophisticated

From terrorism-risk analysis to hurricane modeling, catastrophe modeling and forecasting tools are becoming ever more sophisticated to help carriers make informed decisions.

Major insurers are seeing significant changes, especially in the area of catastrophe modeling. Previously a specialized service conducted by experts, CAT modeling is now being integrated into more mainstream business processes, some industry observers say. The following are some of the new forecasting software products and modeling technology developments announced in the past year.

Risk Management Solutions, based in Newark, Calif., is offering products that can help insurers with the new Supplemental Rating Questionnaire (SRQ) for terrorism risk issued by A.M. Best Company in 2004. In the questionnaire, insurers are asked to show whether they use a terrorism model to quantify their risks and also show their risk accumulation within key target locations. Carriers are also asked to report their largest levels of risk accumulation in any one four-wall structure and model their losses for various terrorism attack scenarios.

“SRQ is a very interesting developmentits main questions are something clients have all been focused on since Sept. 11,” said Paul VanderMarck, executive vice president at RMS. Mr. VanderMarck observed that there has been a lot of attention paid since Sept. 11 to understanding very precisely where insurers place business. The most obvious implication is for the terrorism risk,” he said, “but also there has been recognition that basic concepts of managing the accumulation of exposures are also relevant more broadly.”

Mr. VanderMarck said that using an RMS product called “RiskLink,” insurers can perform a wide range of terrorism-risk analyses. For instance, insurers can identify their “at risk” locations using the RMS target list, which was developed with help from global experts on international terrorism.

With RiskLink, insurers can also quantify multi-line exposure accumulations within buildings, using high-resolution data for urban areas developed with The Sanborn Map Company. RMS said clients can automatically locate a company's top five multi-line exposure accumulations using a search algorithm, and also locate a company's top five multi-line terrorism loss scenarios using 32 different terror attack modes. Other features include the calculation of exposure accumulations and loss estimates for a range of financial bottom linessuch as gross loss, net loss after reinsurance and net loss after TRIA recoveries.

RMS Chief Executive Hemant Shah noted, “The SRQ also recognizes and highlights the fact that property and workers comp losses can be highly correlated in a terrorism event and thus must be managed accordingly.”

Separately, RMS also introduced last year the third generation of its U.S. Hurricane modelit is the first such model to fully represent the physical processes of hurricanes that impact the United States, according to the firm. The upgraded model includes new research into what are called “transitioning storms”storms that behave differently from pure hurricanes, with distinctive implications for loss estimates. RMS said its model works as both a hurricane and a transitioning storm modelthe first of its kind in the CAT modeling industry.

Using improvements in hurricane data and research over the past five years, the RMS U.S. Hurricane model uses ultra-high resolution modeling techniques on storm behavior and building vulnerability. The model, RMS said, uses “an event-generation methodology” to simulate hurricane events throughout their entire life cycle, determining loss correlations for storms that strike more than one region.

RMS also pointed out that the models hurricane windfields are updated constantlycalculated at 15-minute time intervals along the storms path, using factors such as changes in wind direction, upwind surface roughness and intensity as a storm passes. Losses for both wind and surge are calculated and even differentiated based on state-level differences in building code requirements and enforcement.

“Traditionally, hurricane models have been unable to adequately explain historical losses in the Northeast, which were consistently lower than expected using standard hurricane windfields,” said Robert Muir-Wood, managing director of global risk modeling at RMS. He said most hurricanes approaching U.S. northeastern coastal regions are transitioning types. “These storms exhibit different characteristics from hurricanes. The new research at RMS on the impacts of extra-tropical transition will change how companies underwrite and manage hurricane risk, particularly in the northeastern U.S.”

Prices for RMS products vary based on what parts of the applications are licensed. “Our clients pay anywhere from several hundred thousands dollars to several million, depending on the scale of the relationship,” Mr. VanderMarck said.

AIR Worldwide Corporation, a unit of Jersey City, N.J.-based Insurance Services Office Inc., is offering its clients updated services and software for its catastrophe models. Boston-based AIR observed that the application of CAT modeling by major insurers is undergoing a significant change. CAT modeling, the firm said, has previously been a specialized service conducted by experts. But now, it is being incorporated more into mainstream daily business processes and is added into underwriting software applications. “This is happening across multiple lines of business, including workers compensation and life,” according to AIR.

The firms two recent developments introduced to the marketplace in the past year are CATStation and CATools. “CATStation was introduced in the summer of 2003. CATools was introduced in 2002 and was first implemented successfully with ACE Limited at the end of last year,” said Uday Virkud, senior vice president at AIR Worldwide. “These products are new additions to the family of products that we have. CATStation is an extension of our detailed modeling solutions and it is designed for complex, large companies that write multiple lines like property, workers' comp and life.”

Mr. Virkud explained that the Sept. 11 event brought to the forefront for many companies that there were “aggregations and accumulations” of exposures. “For example, a company might be insuring the property of a building but also write workers' comp policies with employees working in that building, and also life insurance policies on top of that. So together, there could be multiple losses,” he said.

Mr. Virkud added that CATStation, designed to work in tandem with AIRs desktop software CLASIC/2, allows more flexible reporting. Losses, he explained, can be reported on an annual aggregate and occurrence basis, by location, portfolio, peril or other user-defined criteria. It also allows insurers to manage every aspect of their catastrophe risk in one application and analyze risks from the portfolio level down to underwriting decisions for individual policies. CATStation uses three modules: Exposure Concentration Analysis, Hazard Analysis and Loss Analysis.

The Exposure Concentration Analysis component provides a geographical analysis of clients existing exposure concentrations and finds out whether they meet clients underwriting guidelines, he said. Clients can analyze, in real time, their approximate exposure to likely terrorist targets. The impact of new policies on an existing portfolio can also be studied, as can overall exposure concentrations at the enterprise level or down to the single multi-location policy level.

The Hazard Analysis module offers critical catastrophe and property risk information. Its designed to provide peril-specific characteristics of the property location, such as storm-surge potential, distance to nearest active fault and Public Protection Classification, as well as risk scores. Clients can simply input the property address to standardize and geographically code the location, AIR said.

The Loss Analysis module helps insurers to assess the catastrophe loss potential of individual risks, policies and portfolios of policies, the company noted. Utilizing detailed exposure information on each location and policy, this module assists in individual risk selection and pricing decisions.

CATStation also allows insurers to combine catastrophe risk information with their corporate guidelines. In combination with AIRs CLASIC/2 software, insurers can run full portfolio-level analysis and study their options for reinsurance, said AIR. CATStation is a browser-based system that can be fully integrated with an insurers in-house underwriting system or operate as a stand-alone product that can be networked anywhere throughout a client's underwriting units.

CATools, the other upgrade introduced by AIR, is an application programming interface that brings a full integration to insurers: direct access to CAT models from the insurers proprietary underwriting system. It can be used to integrate catastrophe modeling into the mainstream business processes within insurance companies.

Mr. Virkud said the pricing for CATStation and CATools is based on “the extent of the implementation of these applications for the company and the size of the company.”

Willis Re, the reinsurance unit of London-based broker Willis Group Holdings, offers a risk-modeling service for its clients that evaluates property-casualty exposures in an underwriting portfolio. First introduced in May 2003, Willis Concentrated Risk Locationsalso called Willis Ctrlhas been developed to identify exposures from perils that are local in their immediate effect, such as terrorism, hail, tornado, fire and flood, but global in a longer-term impact, the firm said. In addition to recognizing exposed locations, total exposures and potential total losses can be estimated, taking into account relevant policy conditions or loss limits.

Clients can specify an area of interest, such as a city, and Willis Ctrl can scan the portfolio for physical concentrations of risk using a “clustering algorithm.” Insurers can also choose a specific point, such as a high-profile building, and study exposures within a specified radiushalf a mile, for example, said Willis. The service can also calculate exposures within a geographic area, such as a county or zip code, across multiple classes and lines of business.

“This is another great opportunity to provide added value service to help clients make better management and underwriting decisions,” said Julie Serakos, senior vice president of catastrophe-management services at Willis Re and the U.S. leader on the Willis Ctrl project.

“The benefits of Willis Ctrl will help our clients effectively evaluate perils and exposures which have traditionally been difficult or expensive to quantify. It provides a new dimension to the understanding of risk in major areas of concern.”

As for the pricing, “this service is primarily available to our reinsurance clients, so the cost is included within our client relationships,” Ms. Serakos said. Currently, there are some 50 primary insurers that are utilizing Willis Ctrl in their relationships with Willis Re.

In the personal-financial analysis area, TransUnion and Convergence Data LLC, both based in Chicago, are offering a service called Loss Improvement Forecasting Tool (LIFT) to personal-line insurers. It works with TransUnion's credit-based insurance software to help insurers better assess risk in underwriting policies.

Kevin Coghill, vice president of TransUnion, said it offers insurers an additional tool to make more informed underwriting decisions. “Just like credit report variables, LIFT data demonstrates a strong correlation to loss. This additional information can lead to underwriting action that will make a solid difference in how the book performs,” Mr. Coghill said.

LIFT data, which is compliant with the Fair Credit Reporting Act, combines information from TeleCheck, a check-verification-and-guarantee company, and Teletrack, a sub-prime consumer-credit database. These data sources, the companies said, contain information on consumer-financial behavior similar to that of credit-reporting companies. LIFT can help protect insurers from adverse risks and reduce losses by further refining credit-based insurance score ranges.

Gregg Antenen, a partner in Convergence Data, said, “Clearly, more and better information can produce better underwriting decisions on more applicants, and that's the purpose of LIFT. Through our relationship with TransUnion, we will be able to deliver LIFT seamlessly into an insurer's application processing system alongside traditional credit data.”

No pricing information was provided.

Reproduced from National Underwriter Property & Casualty/Risk & Benefits Management Edition, February 20, 2004. Copyright 2004 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.

Special Reports /

Cat Modeling, Forecasting Tools More Sophisticated

Related Stories

Recommended For You