Leveraging GIS Analytics for Improved Catastrophe Risk Management

By Roger Grenier

In today's data-driven environment, leading insurance companies are making the capture and use of address-level location information a top priority. Once a location is established, the next step is capturing as much information as possible about that property. Location information is essential for assessing catastrophe risk and for effective catastrophe modeling.

GIS (geographic information systems) technologies provide companies with the ability to combine geographic location with building characteristics to help them enhance exposure data for individual locations. GIS has proven to be a useful tool in spatially identifying where companies have good or poor property-specific building information. That insight helps companies target areas of concern and develop strategies to address potential problems.

In addition to identifying areas to enhance exposure data quality, GIS can be used to analyze the impact of property-specific exposure data in portfolio growth decisions. This enables companies to price the true risk associated with individual properties and identify underperforming areas within a portfolio.

Using GIS Analytics for Exposure Data Quality Assessment
Property characteristics such as building construction, occupancy, roof shape, and foundation type have a significant impact on catastrophe modeling results. The more detailed the property information used, the more refined and reliable the modeling results.

When property-specific building information is not known, it's common for companies to label those fields as "unknown." While catastrophe modeling software will substitute a weighted representation of the common characteristics of buildings in a given area for buildings with unknown values, it is not an acceptable substitute for the actual building information. A recent analysis of a property exposure to earthquake risk in California, for example, showed an order-of-magnitude difference in the AAL (average annual loss) when the building was coded as "unknown" compared with the analysis when the building was coded with known, accurate values.

Considering the importance of building characteristics in catastrophe modeling, it has become necessary for insurers to have high-quality data on risks, particularly when their portfolios have higher-than-average concentrations of properties in at-risk locations. GIS technology offers companies a better understanding of the quality of their exposure data in relation to those high-risk areas.

For example, exposure data validation analyses can be conducted to determine what properties have missing or unreasonable building characteristics. The results are then combined with results of catastrophe models and analyzed using GIS technology, which maps the characteristics in relation to the catastrophe risk. This is as simple as correlations to known catastrophe hazards, such as earthquake fault lines, floodplains, and coastal areas. More sophisticated analyses can account for the intensity and frequency of possible events, thus enabling a more precise correlation between catastrophe risk and exposure data quality. With those GIS tools, companies are able to view missing or inadequate exposure data, helping them focus their efforts to enhance exposure data where it has the largest impact on overall portfolio risk.

Using GIS Analytics to Assess Property Values
Building replacement value is another essential data point in determining catastrophe risk. Catastrophe models estimate loss by assessing vulnerability of a risk based on property characteristics and replacement values before applying policy terms and conditions. If a property's replacement value is understated by as much as 30 percent, the estimated catastrophe loss will be understated as well. Widespread understatement of building values will lead to a significant underestimation of catastrophe risk.

During the underwriting process, most companies estimate replacement costs using component-based replacement-cost estimators. When used properly, those systems provide a reliable estimate of total loss. However, the applications can be influenced for a competitive advantage or other reasons. Furthermore, most companies do not regularly recalculate replacement values. Consequently, the replacement values do not account for changes in the actual building materials or increases in construction costs. Over time, this may lead to a misrepresentation of the actual replacement value.

Using exposure data validation and benchmarking analyses, companies can identify where building replacement values may not be adequate. Data validation reveals whether a replacement value exists for each building in the portfolio and, if it does, whether the estimate is reasonable based on the other characteristics of the building. Through data benchmarking, companies can assess, at the country or state level, how the replacement values in their portfolios compare with that of industry averages.

Once again, GIS technology is used to understand exposure data quality geographically. The results from the data-validation analysis may be mapped to show specific locations where the replacement values are not present or do not make sense.

A benchmarking analysis provides an aggregate assessment of how replacement values compare with industry averages. The results are then analyzed using GIS technology to identify correlations between unreliable replacement values and to determine whether they are consistently under- or overestimated by particular agents or underwriting offices within certain territories. If necessary, action can be taken to address any concerns.

Using GIS Analytics for Portfolio Management
Companies use many geographically based metrics to pinpoint the most likely areas in which to grow profitability or determine where they should reduce exposure. Property-specific building characteristics, essential to catastrophe modeling, are another metric that insurers can incorporate into GIS analyses.

GIS analytics also allows marketers to explore the relationships between different attributes of building data and the impact of those attributes on potential catastrophe loss based on geography. The effects of catastrophe hazards, such as wind and earthquake, on certain building constructions and occupancies are well understood. Knowing the correlation of damage to building characteristics is an invaluable insight when making strategic growth decisions.

A company can conduct a portfolio data-validation analysis to determine the mix of building attributes for individual locations within a portfolio. Data mapping can then illustrate how characteristics are spatially dispersed. For example, if the ­port­folio contains a dangerously high number of light-metal buildings (generally more susceptible to wind damage) in coastal Florida, the company can use the information to adjust underwriting guidelines. Conversely, if the portfolio underrepresents reinforced-concrete construction in an area prone to earthquake, the company may look to write more such properties in high-risk areas for potentially profitable growth.

Conclusion
Catastrophe exposure data plays a critical role in managing overall portfolio risk. Using GIS applications to analyze the results of data-quality assessments enables companies to visualize where they have good or poor exposure data. This helps companies develop more efficient strategies to improve data where necessary. In addition to improving data, companies can use GIS applications to make better decisions on where to grow their portfolios while minimizing overall catastrophe risk.

Roger Grenier, Ph.D., is vice president, AIR Worldwide.