Technology has advanced to the point where an individual risk can be analyzed as part of an automated underwriting process. Furthermore, insurers can minimize the impact on the current underwriting workflow by capturing the data required for sound catastrophe risk management as part of the replacement-cost estimation process.
Many policies today are still underwritten using hazard-based assessments of catastrophe risk. These approaches, including rating territories and distance-to-coast, are insufficient because location is the only aspect of the property’s risk profile considered. The use of these metrics can be counterproductive. A distance-to-coast analysis may suggest underwriting a less attractive risk three miles from the coast, while rejecting a better risk 500 feet from the coast.
A catastrophe risk assessment for an individual property requires more than the location. Catastrophe models evaluate other key variables—such as property-specific building characteristics, replacement value and policy terms—to determine the probability of various levels of loss.
Using catastrophe risk models at the point of underwriting allows underwriters to outmaneuver their competitors by writing policies for properties that hazard-based analyses reject, but are likely better risks.
Incorporating Catastrophe Modeling Into Underwriting
Companies that estimate replacement costs as part of the underwriting process can easily capture the data required for sound catastrophe risk modeling through the underwriting process. This can be done without significant changes to the underwriting workflow or the need to train additional catastrophe modeling experts. Furthermore, by incorporating catastrophe modeling into the underwriting process, portfolio level risk analysis will become more streamlined.
To achieve reliable results, the input data needs to be detailed and high-quality. Fortunately, much of the property-specific data needed for catastrophe risk assessment is also needed to estimate replacement costs, making it easy to collect property information relevant to catastrophe risk during the replacement-cost estimation process.
Property-specific information including address, construction type, occupancy type, year built, and number of stories are key data points for catastrophe modeling and are captured by a majority of insurers for replacement-cost estimates.
Additional data elements often collected to estimate repair costs—including roof shape, roof cover, exterior wall type, and building shape—are also important for generating reliable catastrophe loss estimates.
Another key reason to connect the catastrophe modeling and replacement-cost estimation processes is the availability of a current replacement-cost estimate to use in the modeling analysis. Catastrophe models estimate loss by assessing vulnerability of a risk based on property characteristics and replacement value before applying policy terms and conditions.
For example, if a property’s replacement value is understated by 30 percent, the estimated catastrophe loss will be understated by at least that much. At the portfolio-level, widespread understatement of building values lead to significantly underestimated catastrophe risk. Therefore, accurate replacement values are essential for obtaining accurate catastrophe loss estimates.
Leveraging the property-specific data captured during the replacement-cost estimation process provides value for portfolio-level catastrophe modeling, even if catastrophe risk is not analyzed during the underwriting process.
Catastrophe-modeling data extracted from insurers’ policy management systems can be enhanced, augmented, or replaced with data available in the replacement-cost estimator. By extracting property data directly from the replacement-cost estimator, insurers can be more confident their analyses are based on detailed and reliable building information for each risk. Furthermore, before the data is extracted for catastrophe modeling, replacement-cost estimates for all properties can be recalculated to ensure they are based on current building costs.
For properties that have been in a portfolio for a number of years, it is advisable that the data be analyzed to ensure it is current, reliable, and detailed. New technology has recently been developed to validate and score catastrophe exposure data. If the data is deemed unreliable, it can be augmented using databases of property-specific information based on physical property inspections and public records data.
A full assessment of catastrophe risk at the point of underwriting will not only enable the underwriter to make better risk selection decisions, but will also help to improve the overall portfolio as it grows. Companies will be able to manage exposure concentrations in real time based on catastrophe loss potential, not insured values, providing a more precise view of the risk.
A company that builds its portfolio with risks vetted by a catastrophe model during underwriting will also enjoy other benefits. It will be able to better plan its reinsurance needs, as it will have a more timely and accurate view of the risk. Also, a company with such robust catastrophe risk management practices will be in an excellent position to negotiate the best reinsurance pricing.
Rating agencies, including A.M. Best and Fitch, are putting greater scrutiny on catastrophe risk management practices and the exposure data use for the analyses. Insurers that take a proactive approach to catastrophe risk management will be well-positioned during a ratings review.
Preparing for the Future
Florida will always face a significant hurricane threat. However, the ability to effectively manage the risk is constantly evolving. From a workflow perspective, probabilistic catastrophe analysis has progressed to the point that it can be incorporated into a fully automated underwriting environment.
A full assessment of catastrophe risk at the point of underwriting will enable the underwriter to make better risk selection decisions, and is also an important part of a robust and comprehensive approach to managing catastrophe risk.