|
This article originally appeared in the August 2008 edition of ISO Review.
Feature Story:
Reliable Catastrophe Modeling Results Require Reliable Exposure Data
by George Davis, FCAS, MAAA, Vice President, AIR Worldwide
A significant challenge confronting today’s insurance and reinsurance industry is data quality, specifically the quality and completeness of property-specific building characteristics and the reliability of replacement-cost estimates. In the catastrophe modeling sector in particular, data quality is lacking — an issue brought to the forefront by the hurricane events of 2004 and 2005. A lack of reliable exposure data can have a negative impact on catastrophe risk management. Insurers need to take a close look at the exposure information they collect for catastrophe analysis and, if necessary, put new processes in place to improve the data.
Insurers that accomplish this goal will enhance the reliability of their catastrophe modeling, improve underwriting decisions, better manage exposure accumulations, and more reliably assess reinsurance needs. Additionally, now that the industry is placing more emphasis on collecting better exposure data and catastrophe modeling assessment, insurers that provide high-quality exposure data and can display robust insurance-to-value practices will place themselves in a better position to structure reinsurance terms. Furthermore, rating agencies are placing more emphasis on the collection of exposure data when assessing the catastrophe risk management practices of insurers.
Much of the property data used for replacement-cost estimation can also be used for catastrophe risk analysis
Key Building Blocks
The key building blocks of high-quality exposure data are location, property-specific risk characteristics, and replacement values.
Location
Insurers should focus on capturing geocodes for each property they insure or at least high-quality address data that can be translated into the geocode. The exact latitude and longitude of the property are often critical to effective catastrophe modeling. In the absence of an exact geocode, insurers may assess the risk of the property based on the ZIP-code centroid or city centroid. This could result in the model positioning the risk further away from the hazard (e.g., coastline or fault line) than it is in reality, potentially leading to an underestimation of the risk.
Property-specific risk characteristics
Property-specific risk characteristics can have a significant impact on catastrophe modeling results. To demonstrate the point, AIR conducted a sensitivity analysis assessing the impact of unknown property data on a commercial and residential portfolio of Florida properties in comparison to an analysis where the exposure information was known. The results are clear. Catastrophe modeling becomes more realistic as more property-specific information is added.
Replacement values
Since catastrophe models estimate loss by assessing vulnerability of a risk based on property characteristics and replacement value before applying policy terms and conditions, accurate replacement values are essential for obtaining accurate catastrophe loss estimates. If a property’s replacement value is understated by 30 percent, the estimated catastrophe loss will be understated by at least that much. Widespread understatement of building values leads to a significant underestimation of catastrophe risk.
FEMA/Michael Rieger
Strategies for Improving Exposure Data
Strategies are available today for insurers to assess and improve the quality of their exposure data at the point of underwriting and for policies already in their portfolios.
At the Point of Underwriting
Fortunately, most of the exposure information necessary for catastrophe risk management is collected during the underwriting process, since, in many cases, it’s the same data used to estimate replacement costs. This data can be collected by agents and can often be provided by property-specific databases, such as ISO PushPin and SPI Plus®, to make the data-collection process more efficient.
The challenge for some insurers is capturing the data and replacement-cost estimate in the policy-management system for use in catastrophe modeling. 360Value™, a replacement-cost estimator developed by ISO companies Xactware and AIR Worldwide, is available to facilitate this process today. Insurers that make use of property data collected during underwriting will go a long way toward enhancing catastrophe analysis results.
Policies Already on the Books
Insurers have access to services to assess data completeness and data quality by comparing portfolio data with exposure data embedded in detailed industry exposure databases. In cases where company-collected data is incomplete or of insufficient quality, data can be augmented with reliable exposure data provided by AIR.
To keep pace with insurance to value, insurers should revalue their policies using the most recent information about the property and the latest building-cost data. For a seamless process that ensures replacement values are current every year, insurers can incorporate an automated renewal process to revalue annually. Additionally, insurers can conduct portfolio analyses where insurance to value can be assessed at both the portfolio and policy levels.
The Future Looks Bright
Improving exposure-data quality for catastrophe risk management is a continual challenge. Insurers need to develop processes to collect and maintain exposure data that complements their particular workflow. Solutions to help insurers collect and maintain exposure data and assess the quality of the data they have already collected are available today. AIR is developing additional exposure-data analytics and compiling an enhanced database of property-specific information for millions of residential and commercial properties in the United States. Shortly, these capabilities will be available directly through the catastrophe modeling system and other products.
»Return to home page
|
Add a Comment 
On-demand webinar on this topic