Roof experience key for insurers seeking to mitigate hail risk

The traditional methods of assessing hail risk are no longer sufficient.

Insurers must now consider both aerial imagery-derived insights as well as a property’s recent hail experience to overcome gaps in accurately evaluating and mitigating hail risk. (Credit: Vladimir Polikarpov/Adobe Stock)

In the rapidly evolving peril landscape, the increasing prevalence of hail-related claims, exacerbated by significant exposure growth and possibly even changing weather patterns, is a critical concern.

The number of hail events this year has already increased by 48% over the number from 2022, according to the National Oceanic and Atmospheric Administration. The growing number of losses — wind and hail accounted for 39.4% of all losses in 2021, according to the Insurance Information Institute — has made clear that the traditional methods of assessing hail risk are no longer sufficient. As it is not feasible to completely avoid the peril altogether, insurance carriers must now consider both aerial imagery-derived insights as well as a property’s recent hail experience to overcome gaps in accurately evaluating and mitigating this growing risk.

The first piece of the puzzle is understanding a property’s vulnerability to hail by leveraging imagery-based insights. Geospatial imagery analytics, like those derived from AI technology like computer vision, can analyze high-resolution aerial imagery to derive structure-specific vulnerability characteristics such as roof shape, size, composition, complexity, condition and more. Hail impacts can vary from property to property, so it is essential for insurers to incorporate these vulnerability insights into their hail risk assessments as they can affect both the likelihood and severity of an eventual claim.

However, imagery-based analytics alone fall just short in revealing the complete risk picture. By relying solely on imagery analytics and regional hazard scores, carriers could miss pre-existing roof damage because hail impacts cannot be seen in even the best high-resolution imagery available today. Homeowners may not know about this pre-existing damage, either. It’s important to understand that prior minor hail incidents, which might not have led to claims or obvious visible damage, cumulatively degrade roof materials like asphalt shingles. Even the damage from larger hail incidents is often undetectable from imagery. Consequently, these homes may suffer unexpected severe damage when hit by larger hailstorms. Carriers may be “buying” hail claims more often than they think. 

The missing piece for insurers is a property’s recent hail experience. This includes leveraging forensic weather analytics, including data about the number of recent hail events and their maximum size, to help shed light on this critical blind spot in a home’s hail risk assessment. Our research at CAPE Analytics shows that 60% of homes with significant recent hailstorm history (i.e., multiple events and/or single events with hail sizes over 1.75 inches) still display good or excellent roof conditions in imagery-based evaluations.

Carriers cannot rely on just any weather data, however. Hail is notorious for being difficult to track, and — as shown by field research conducted by the Institute for Business and Home Safety — two homes a mile apart may have very different experiences in hail size and concentrations. In addition, historical hail event data has known population bias given the human reporting requirement for publicly available datasets. What carriers need are forensic weather analytics — the next generation of analytics that can determine hail history on a property-by-property basis.

Recent hail risk research from CAPE underscores the concern that an extensive hail experience can cause. We found that properties with a higher frequency and severity of hailstorms in the previous two years are 50% more likely to have future claims due to undetected damage.

These findings are a call to action for insurers: Having a complete understanding of a home’s hail risk by taking into account its recent experience is essential to effectively manage this peril and provide homeowners with accurate coverage. Approximately one-third of roofs considered in good or excellent condition by imagery-based analytics alone would trigger additional review after considering their recent hail experience.

Knowing a roof’s hail experience enables carriers to deploy proactive measures, such as reaching out to homeowners to verify recent roof replacements following the most recent image or hail event or conducting thorough inspections before policy binding. This key insight can also help insurers more proactively manage potential claims, establish more refined risk segmentation, and mitigate financial risks when reviewing an existing book of business.

With the growing amount of pre-existing damage nationwide, knowledge of a property’s hail experience will be essential. Canopy Weather estimates that through the first three quarters of 2023, there was already $37 billion in pre-existing hail damage on carriers’ books — damage that will persist for several years.

What does this mean for insurers and underwriters looking to integrate hail risk into their workflows? Integrating historical forensic weather analytics with current imagery-based analytics allows for a more comprehensive view of a property’s hail risk and susceptibility to future claims. By embracing innovative solutions that incorporate a property’s hail history, the industry can more accurately assess and mitigate these risks, ultimately leading to better outcomes for both insurers and policyholders.

Kevin Van Leer

Kevin Van Leer is vice president of Client Success at CAPE Analytics. In this role, he heads the Client Success team at CAPE Analytics, with a focus on maximizing the value delivered to clients from CAPE’s imagery-derived insights to enable effective risk differentiation and more efficient underwriting workflows. Prior to CAPE, Kevin spent five years at RMS, driving the release of several climate-peril catastrophe models, including wildfire, hurricane, and severe convective storm, alongside location-specific vulnerability analytics.

These opinions are the author’s own.

See also: