Over the past decade, there has been an explosion in the number of digital analytical tools available to underwriters and brokers. The insurance industry has always been data-driven, but modern catastrophe risk modeling allows us to use data to create a clearer picture of property risks. Adoption of these models has been almost universal and cements their prominent role in insurance and reinsurance analysis and pricing.
It is easy to understand why—some of the possibilities are impressive. The recently released Touchstone 2.0 from AIR Worldwide makes it much easier to model severe thunderstorms and tornadoes. Corelogic's RiskMeter product is available online and looks at 30 natural hazards, from physical hazards to windstorm risk. Another company called Intermap, a provider of geospatial solutions, recently let me beta test their Insite Pro product, which allows users to leverage dependable ground elevations to better qualify flood risk and also features a user-friendly mapping interface.
These tools, alongside trends in metadata usage, have opened up possibilities for understanding risks. However, any mistakes made in initial data collection can have ripples throughout the process, ultimately affecting pricing.
Catastrophe risk modeling must always start with one crucial piece of information: the correct address. Any model is only as good as the data you supply, and your data set will always be insufficient if you do not provide an accurate, complete address.
This sounds profoundly basic. Of course you need a specific address to understand the exposures faced by a particular risk, right? But it is not uncommon for brokers and underwriters to work with incomplete or inaccurate address data. Why, and why is this such a big deal?
Sometimes the sheer size of a risk makes it difficult to get specific address data. When a large company is trying to insure thousands of locations, getting this data into a model-friendly schedule is a laborious task. In other instances, geographically large properties may contain a range of addresses, which could hurt geocoding accuracy.
Local quirks might create big problems for analytical results. For example, people who live on small highways often refer to their street name differently than their municipality; they may call it Main Street while their town calls it Route 1, or vice versa. Many streets share similar spelling or pronunciations (e.g., Karat Street and Carrot Street). Discrepancies like this can also snag brokers and underwriters who are modeling a risk.
If the risk modeling programs are not given sufficient address data, they will default to "postal code centroid." This essentially identifies the location of the risk as the middle of the zip code in which it is located. This stems from how Geographic Information System (GIS) programs handle incomplete address information.
To use Intermap as an example, if provided with only a ZIP code, the model will not make the most of its ability to show slope difference. Within one ZIP code, some properties could be 200 feet above sea level while others are in a 100-year flood zone. Using the incomplete or inaccurate data given, the analytical programs will make "educated guesses" based on damage functions. Half a mile—or even a block—can make a difference when assessing and pricing an insured's flood zone determination.
Accurate address information is not only important when it comes to the peril of inland flood and coastal storm surge associated with a windstorm. When modeling for earthquakes, improper location data can lead to an exposure being placed in the wrong seismic zone or show the improper distance to a fault line.
To get the most useful results during the analytical stage of assessing risk, brokers should educate their clients on the importance of precise data. The insured should understand it is best not to leave anything to chance when assessing their risk. And by providing imprecise address information, much is left to chance.
Accurately modeling a risk provides the analytics you need to better understand your risks. The more precise the data obtained, the more confidence the insured, broker and underwriter can have in the model's results and pricing of a policy.
Technological innovation is now a key building block of our industry, and we are just beginning to see the new possibilities afforded by more powerful analytical tools and data. Yet as we innovate, we cannot forget our foundation: quality data. Execute the basics well.
© Arc, All Rights Reserved. Request academic re-use from www.copyright.com. All other uses, submit a request to TMSalesOperations@arc-network.com. For more information visit Asset & Logo Licensing.