According to data from Verisk's A-PLUS™ property database, 2013 industry hail loss was $4.9 billion, the lowest industry loss since 2007 and well below the industry average loss of $6.4 billion between 2008 and 2012.
However, insurers shouldn't rest easy, as there isn't yet conclusive to suggest whether hail is becoming less frequent — or more.
Managing losses from future hail events
There are three primary strategies insurers can implement to manage the risk of future hail losses. First, complement traditional rating methodologies with the use of weather-based predictive analytics and roof characteristics when developing an appropriate hail premium. Second, focus underwriting efforts on ensuring the integrity of rating characteristics (such as replacement cost) and on other data unavailable during quote or where rating data is too thin to be reliable. Third, lessen your exposure through cost-sharing opportunities with consumers, such as increased deductibles and depreciated roof coverage.
Hail risk is likely to continue to escalate and challenge property underwriters for two reasons. First, exposure is increasing in hail-prone areas because of general population growth and climbing reconstruction costs. Second, in geographies with few recent hail losses but the potential for a sizable hailstorm, many roofs will require replacement at once. That, combined with roofer and materials demand surge, could produce heavy losses and significant earnings volatility.
The challenge is determining property-level exposure to future hail events and each risk's proximity to them. You can effectively address such a challenge through predictive analytics based on the innovative use of high-quality, long-baseline hail-related weather data informed by claims data and analytics.
Characteristics of a roof, including age, material, and construction, can all affect how a roof performs when struck by hail. Among other findings, a report by the Roofing Industry Committee on Weather Issues determined:
- damage is less likely to occur on an asphalt shingle roof less than ten years old
- damage to metal roofs is generally cosmetic, rather than functional
- roof cover built on a solid substrate fared better than an unsupported roof cover
Data on roof age, material, and construction can provide unprecedented insight for underwriting decisions. Current sources of data include homeowners, property tax assessors, building permits, and property inspections. Each source has advantages, but an overall understanding of the roof risk is essential to hail underwriting.
An emerging source of property-specific roof information is recent claims estimates. Every time insurance repair contractors develop an estimate to repair the roof of a house, they create detailed records of the roof, including size, material, slope, shape, and more. In addition to roof characteristics, claims estimates can also provide the effective roof age and roof replacement cost. The claims estimate can capture those details by address, and insurers can use the data to help guide underwriting decisions.
Identifying unrepaired hail damage
Hail losses between 2008 and 2012 were unprecedented, as shown in the chart below. According to the A-PLUS property database of historical claims, policyholders filed more than 4.5 million hail claims at a cost of $32.1 billion. While those losses were extremely high, Verisk Climate estimates that damaging hail has likely affected as many as 10 million households in the same time frame, leaving the potential for unclaimed damage at more than 5 million locations.
There are several reasons why property owners may not file a claim immediately after an event. They may be unaware of hail damage, which can be difficult to spot from the ground. Even if damage is known, property owners may choose not to repair, thinking it's not worth the out-of-pocket expense or inconvenience. That choice may cause more trouble down the road if unrepaired hail damage causes a roof breach or weakens the roof and another storm strikes the property.
During underwriting, knowing the likelihood of preexisting damage without a known repair can assist an insurer in maintaining profitability.
After a hailstorm, it's hard to miss the houses with lawn signs for roofing companies repairing the damage. In most cases, each of those properties also has an insurance claim. One strategy insurers can use to assess the likelihood of unrepaired hail damage is the analysis of claims loss histories for properties in proximity to the one under consideration for underwriting.
You can run analyses to identify hail claims within a defined distance and time frame based on the applicant's address. Multiple claims with a common cause of loss signal an elevated risk for unrepaired damage.
Similar to historical claims information, analytics based on historical weather events — particularly, recent events — can provide tremendous insight into the likelihood of preexisting hail damage. High-resolution weather history data gives insurers information about each storm that may have affected the property, including date of occurrence, hail size, and duration. The data can be the basis of a score representing the likelihood that damaging hail affected the property.
Insurers can use each of those strategies independently or combined to give a solid indication that they should take action to ensure no prior damage exists.
A new underwriting approach
The property loss landscape has changed, and hail risk is now a larger part of total losses than it has been historically. Verisk is committed to developing new data and analytics to assist you in making good decisions, being responsive to environmental changes, and managing your book of business.
Photo Source: NOAA National Weather Service. Sources for Hail Facts infographic: www.xactware.com/solutions/industry-trend-reports; analysis by Verisk Insurance Solutions – Underwriting, A-PLUS; Xactware analysis of claims loss estimates between 2008 and 2012; LOCATION® Hail Damage Score analysis by Atmospheric and Environmental Research (AER); http://www.spc.noaa.gov/climo/online/monthly/newm.html#2012.