Investigating Increase in Hail Claims
Over the last decade, we’ve noted a perception in the insurance marketplace that hail events have increased and had a strong effect on property claims in the U.S. We thought it would be interesting to explore the validity of this perception and its nature. We also wondered about the possible reasons - or biases - driving this perception. Here we share our observations.
We began our exploration by doing a comparative study of research material on frequency of hail events. Most of the widely-cited research papers, using 30 to 100 years of hail events for their analyses, suggest that there is no systematic increase in a countrywide trend for frequency of hail events. While we could not find any statistical research on trends in severity of hail events with severity quantified by hail size, we did find reports discussing trends in hail severity as quantified by associated claims. One limitation of the severity research is that the latest year for reported data is 2006. Only one of the research papers stated that the major hail losses have increased over time, but the same researcher also wrote in a different paper that the total of all losses has not increased over time.1
To conduct our own review of severity, we utilized large losses through ISO’s Property Claim Service unit (PCS) and performed a statistical analysis of this PCS data to test for the severity trend. We adjusted the claims for inflation and exposure across the Personal, Commercial and Auto lines. Our analysis of the results revealed no change in trend observed in the claims for any one of these individual lines or all lines combined.
In addition to looking at research papers and PCS catastrophe loss data, we also studied the media reports that are propagating the perception of increased hail claims. Much of the discussion in the media on increase in hail claims can probably be traced back to two reports: “Property Hail Claims in the United States: 2000-2013” by Verisk and “Keeping Pace With Texas Hail Claim ‘Case-Runners’” by Zelle Hofmann Voelbel & Mason LLP.2
The Verisk report does not mention any adjustment to claims for inflation or increase in exposure, and it goes back only 13 years - a short timeframe to comment on a trend for severe weather patterns. We would advocate that hail and tornado patterns should be viewed over at least 25 years and preferably 50 years.
The updated version of the Verisk report states, “2014 is on its way to being a better year for hail losses than recent years.”3 According to the original report, “losses in 2013 were the lowest of the past six years but still 50% higher than the worst year from 2000 through 2007.” The recent downward trend of 2013 and 2014 makes the years 2006 to 2012 as outliers, though this conclusion is statistically non-rigorous.
The “Texas Case-Runners” report identifies “the practice of case-running by public adjusters and other third parties involving themselves in the property insurance claim process” as the cause of claims increase. The report focuses on Hidalgo County, TX which has seen 35% of hail claims going into litigation. We couldn’t find any other county in the U.S. where this particular issue is causing an abrupt increase in hail claims. Litigation after large catastrophes has become more apparent at higher dollar levels across Texas, most notably with Hurricane Rita in 2005 and perhaps even more prominently after 2008’s Hurricane Ike.
In recent developments there are two pending bills in the Texas legislature that are attempting to address this looming crisis, and among the proposed measures is one that would ease an 18% annual penalty on insurance companies that don’t pay claims within 60 days. On the trial front, the first completed case against an insurer stemming from one of the 2012 Hidalgo County hailstorms that came to a jury decision has been found in favor of the insurance company. So there is recognition of this issue, and it may be resolved in the near future.
Solar Panel Losses
One rapidly growing technology that would appear to be particularly susceptible to hail damage is solar panel arrays for residential and commercial rooftops. While the numbers are relatively small right now, the growth in usage is expected to be significant over the next few decades - and therefore is a looming insurance concern.
As of the end of 2014, the total number of homes with rooftop solar panel systems across the U.S. was only 600,000.4 In addition, we found minimal overlap between the states with high hail susceptibility and those with the greatest number of solar panel installations. So it seems that residential solar panels will have a low to negligible impact on countrywide claims due to hail.
U.S Solar Panel Installations by State, 2014
Source: seia.org, U.S. Solar Market Insight Report, 2014 Year In Review, Executive Summary
© 2015, Greentech Media, Inc. (GTM Research) and the Solar Energy Industries Association (SEIA). Reprinted with permission.
For insurance coverage anyway, if there is damage, coverage depends on whether the solar panels are owned by the homeowner or installed and operated through a lease or a purchase power agreement (PPA) from solar service companies which is not allowed by all states.
Insurance coverage for self-owned panels differs from one insurance company to another. Some companies don’t require additional insurance, while others require limits to be increased (in some cases up to a minimum of $300,000) or require the owner to purchase an additional cover. The panels that are acquired through a lease or a PPA are insured by the solar panel service provider that owns the solar panels.
Hail - A Reinsurance Underwriter’s Perspective
Hail is another one of the many catastrophe perils that impacts our clients with frequency and occasional severity. A very intense, concentrated event can cause a disproportionate distribution of losses to a cedant’s portfolio. Several years ago, our clients in Europe suffered a very localized event that created over $1 billion of industry insured loss in less than 15 minutes.
Much work is being done on the probabilistic modeling of Severe Convective and tornado/hailstorm perils but they are still less robust than the current market offerings for hurricane models. In our previous Tornado article we referenced some of the major limitations, including less discrete exposure input data and questions over frequency assumptions built into the models. Companies are adding a layer of analysis on a deterministic basis by analyzing the impact of various event characteristics on areas of highly concentrated exposures and total insurable values (TIVs) to come up with another view of “how bad can bad be?”
On the primary underwriting and rating side of the coin, companies have been steadily refining their approaches to coverage options and limitations for roofs based on materials, age of roof, prior damage limitations, cosmetic damages and instituting firm guidelines on when they will offer Replacement Cost versus Actual Cash Value settlement options. Deductibles have been addressed by many companies, including options that are based on percentages of TIV versus fixed dollar variants. There is much activity concerning the evaluation of roofs in the underwriting process with improved sources of data available from building permits and aerial and satellite imagery. We can envision visual inspection via drones becoming a standard practice in the not-too-distant future.
Lastly in the primary pricing venue, the increasing proliferation of Predictive Models is bringing hail into greater focus as part of the “by-peril” rating methodology, adding more granularity in the pricing of individual risks.
Based on our analysis, we do not believe that the recent observed increase in hail claims is reflective of any systematic increase that will continue into the near future. The number of hail incidents hasn’t really changed, so in looking past the short term trend in favor of the long term, we would lean toward a flat long-term trend. Year-to-year changes in hail event frequency and/or severity can create two possible undesired scenarios: an unprofitable result when trends outpace expectations or, conversely, a competitive threat when other carriers perceive a reduction in trend and swiftly act upon it through preferred pricing and coverage.
With the average catastrophe major model version change occurring every five years, companies are individually challenged with understanding, monitoring and evaluating change in trend. We can help with that.
For More Information
For information and resources from the Insurance Institute for Business & Home Safety - including its 2015 Hail Field Study - visit disastersafety.org. Or follow them on Twitter at https://twitter.com/IBHSHailStudy.