The Future of Weather Analytics

It's no secret that 2011 was a record year for catastrophe loss: Carriers incurred $380 billion in losses compared to the previous high of $220 billion in 2005. In addition to unprecedented triple-digit heat and drought, the number of severe natural catastrophes globally was unparalleled.

2012 has not yet proven to be quite as extreme as 2011, but significant global insured losses are already accruing. According to Aon Benfield, losses for the month of April are close to $1 billion, with a significant contribution from U.S. storms and twisters. May is typically the most active tornado month of the year in the U.S., and the potential for greater losses like those seen in 2011 still exists.

With the possibility that thunderstorms, tornadoes and hail are either increasing in frequency and severity or migrating locations in which they most commonly occur, carriers and reinsurers are expressing the need to become more aware of weather trends.

In recent years, carriers have focused significant effort and expense on retooling their claims operations to become more responsive, accurate and efficient, as well as investing in new technologies including claims systems, interactive call centers, tracking and scheduling systems, and mobile applications. The greatest opportunity to gain or lose customer loyalty occurs during a claim, and carriers' efforts to modernize have been relatively successful in increasing customer satisfaction.

But customer loyalty is not the sole concern. Carriers typically spend 3-5 percent of every premium dollar in adjustment expense, and getting a response team on the ground can be challenging and costly during the time immediately following a storm. The role of the field adjuster is still critical in surveying the damage, scoping out the loss, and taking photographs and measurements on site—but it is a costly and time-consuming process to complete this work. The more of this claims work that can be done from the desk, the more efficiently and cost-effectively claims can be processed.

The good news is that there is an enormous opportunity to fast-track the process of accurately identifying the number and severity of potential claims, thus reducing processing costs. Sophisticated weather and claims analytics tools are now available that can minimize the manual efforts associated with determining the extent of weather-related losses.

Advanced awareness of location-specific weather patterns can allow carriers to preposition staff and prepare resources—and these tools can accurately map the areas in which to deploy adjusters, thereby saving time and expense. Furthermore, by clearly delineating the area of impact and the severity of the storm, anomalous loss patterns can be more clearly identified by both size and location; claims that fall outside of the range of probability may be the result of a prior storm or even fraud.

There are several trends in the most cutting-edge weather technologies and analytical tools that will bring down costs and increase accuracy of claims processing.  But in order for insurers to effectively automate some of their claims estimations, two important factors must be considered:

• For weather and environmental data to have value, it must be highly accurate; low-resolution data will be insufficient to make useful predictions. Interactive high-resolution maps with detailed shape-mapping of hail size and location can provide much greater insight into the probable magnitude of loss from a weather event. By leveraging these maps, insurers can not only identify which properties were likely impacted by a storm but also estimate the severity of the potential damage.

• In order to accomplish the most accurate estimation of damages, yesterday's toolbox of weather radars alone cannot address the issue: Weather data must be synthesized from multiple platforms. Lightning, hail, wind and tornado damage frequently occur simultaneously and can have regional synergistic effects. One-inch hail may cause minimal damages to roofs in one area, but combined with high wind shear, the severity of damages can be increased and the mapping of that damage may be directionally skewed. Multiple platforms from computer models to intelligence-grade satellite processing are now available to assist in depicting a more complete picture.

A platform that integrates data sets from the carrier can also be quite valuable in accurately deriving a range of probable loss by property. By overlaying policy-in-force data onto weather tools, carriers can more quickly estimate loss reserves by calculating the property values and deductibles of the policy against the likelihood and severity of the event.

Renee Beauford, vice president of claims for Hastings Mutual Insurance Co., has found significant benefit in leveraging these next-generation weather analytics tools. "We use AER Respond on a daily basis, overlaying high-resolution weather data with our claims and [policies in force].  Ingesting their near-real-time data gives us an edge on deploying the right adjusting resources to the right problems quickly and accurately."

Integration of tools such as these into existing workflows can streamline and standardize the process of analysis. Valuation tools, like those used by appraisers or contractors to determine building and replacement costs, can also be incorporated into calculations for even greater accuracy. Ultimately, the goal is for the insurer to be in a position to anticipate and quickly respond to any claim initiated by the policyholder.

As weather- and Earth-observation companies acquire more historical-cause-of-loss data over time, the future of weather analytics is moving from the claims to the underwriting side, managing risk profiles based on statistics for more granular regions rather than using generalized maps that may not take the individual variations between properties into consideration. Insurers may not be able to control the frequency of catastrophes, but they can certainly control their losses through better claims management—and ultimately, through better underwriting methodologies.