Inside the Hail Lab Confronting a $30 Billion Challenge

The risks that prove most costly are not always the ones that dominate the headlines.

While businesses and governments have poured billions into hurricanes, wildfires, and floods — disasters that command front pages and monopolize political attention — other so-called ‘secondary perils’ have steadily increased in frequency and intensity, inflicting human and financial losses that are now rivalling those of their more high-profile counterparts.

This is particularly true of severe convective storms (SCS) and their most dangerous payload: hail. In the US, an inflection point appears to have been reached. A single catastrophic hail event in the US can now inflict roughly the same financial damage as a Category 4 hurricane, according to financial data company Cotality. A storm with a 2% annual probability of occurrence — a 1-in-50-year event — could generate US$30bn in insured losses alone. For comparison, Hurricane Milton in 2024 produced between US$6-11bn in insured losses, according to catastrophe risk modeling firm Verisk.

And yet understanding of SCS and hail risks, and the industry’s ability to predict their financial impacts, remains limited. “If you compare perils, a catastrophe [‘cat’] model for hurricane risk I would say is probably a ‘teenager looking to graduate high school,’” says Victor Gensini, a professor at Northern Illinois University’s Department of Earth, Atmosphere and Environment. “Cat models for SCS are like literally just getting out of diapers.”

Gensini also directs the Center for Interdisciplinary Research on Convective Storms (CIRCS), a collaboration launched last September between Northern Illinois University, the University of Wisconsin-Madison, and 13 insurance and risk modeling industry partners. Its mission: to produce data and research that insurers and catastrophe modelers can actually use to improve their understanding of SCS risks — before these storms outpace the tools designed to price them.

That knowledge is sorely needed, given the Trump administration’s retreat from climate science and data. While CIRCS itself received US$1.5mn from the US National Science Foundation to get started, this is for a five year period. Further support is being provided by the private partners, each of which pays an annual membership fee. The model — focused research co-funded by universities and commercial firms — may well prove replicable elsewhere.

“Industry will really guide us in terms of their needs, and where they would like to see their money best spent in terms of research and development. We’re sort of building the plane and flying it at the same time,” says Gensini. 

Measuring the current and potential future financial impact of hailstorms is difficult because their destructive power is only obvious when they hit populated areas. When hail falls on undeveloped land, no economic damage is done, and therefore no financial data can be gathered. Moreover, the thin historical record of SCS makes it hard to set a baseline for the frequency and severity of hail-bearing storms compared to other perils, or establish a starting point for forward-looking projections.

“Good hurricane data dates back to the early 1900s. For SCS, we started keeping good records in the 1950s — and that was for tornadoes. For hail, the really good records don’t start until the modernization of the National Weather Service in the US, which happened in the mid‑90s,” explains Gensini.

To address this gap, he and his colleagues at Northern Illinois University have pioneered post-storm field expeditions to gather first-hand data on the size and density of fallen ice. But this approach is not scalable. “Reporting is very, very difficult because reports only come in where there’s people. [That is why we’re asking] how can we use remote sensing, either satellite or radar or other techniques, to try to create synthetic reports or understanding of where observations might be deficient,” says Gensini.

Figuring out how cutting-edge technologies and techniques can be harnessed for hail risk assessment is a core purpose of the Center — and a key draw for its commercial partners, which include Allstate, American Family Insurance, Demex, Guy Carpenter, Moody's, and Verisk, among others.

“One of the reasons you see more of these types of initiatives is that the perils are happening, the losses are accumulating, and there’s a general sense within the insurance industry … that the pricing of risk cannot keep up with the losses that are occurring,” says Matt Coleman, President, Reinsurance at Demex, a specialized reinsurer and insurance technology company. “So there is a demand coming from industry for this type of research that can be voted upon and packaged in a way where there are clear deliverables defined and clear timelines,” he adds.

The data reinforces Coleman’s reasoning. Swiss Re analysis shows that insured losses from SCS are growing at around 7% annually in North America. Hail damage amounts to between US$8bn and US$14bn per year in the US alone, according to the non-profit Climate Central. The National Oceanic and Atmospheric Administration (NOAA) counted over 5,400 hail events in the US last year, with around 800 involving large hail of two or more inches — the kind that can really pack a punch. 

High losses, year after year, are putting carriers under strain. “The evolving dynamics of thunderstorm risk over the past five to ten years have disrupted pricing and capacity within the insurance sector,” says Josh Darr, Head of North America Peril Advisory for Guy Carpenter and Vice Chair of CIRCS. “Our clients, primarily insurance companies, have experienced several years of unprofitability due to exposure to severe thunderstorms.”

Several factors are driving this trend. Climate change is one. Hail is formed when rain droplets are sucked high into the atmosphere by thunderstorm updrafts and freeze in the cold air above. They then grow by merging with supercooled liquid water that hardens onto their surface. A hotter atmosphere may produce larger hailstones: higher temperatures can intensify updrafts, pulling more droplets upward and holding them in the supercooled air for longer.

In the US, meteorological research points to an eastward shift in hail risk, which may also be contributing to insurers’ woes. From 1979 to 2017, conditions favorable to large hail stones expanded across much of the central and eastern US, with the steepest increases in the Midwest. This means parts of the country that used to be spared frequent and damaging hail strikes are now encountering them much more often.

“The shift eastward, along with the long-term growth of all cities [and] the expanding bullseye of the urban footprint just makes for larger losses per event when you have that exposure,” says Charles Jackson, Assistant Vice President, Atmospheric Sciences, at Verisk. Indeed, Swiss Re says that growing exposure and vulnerability to storm risks — and not changes to the hazard itself — are driving the majority of insured loss growth. The ageing, unfortified condition of the median US property is another reason hail is extracting an increasing toll.

Verisk is one of a small group of catastrophe modeling firms that supply insurers and other institutions with estimates of potential SCS losses. Its membership of CIRCS is motivated in part by access to high-resolution climate data showing how warmer conditions could intensify hailstorms at a local level.

Mercury Insurance — a California-based property and casualty insurer — is another member of the Center. Head of Climate and Catastrophe Science Stephen Bennett sees his company’s participation as a way to outsource the time, cost, and effort of improving its own understanding of hail risk.

“The hard science: how does the storm itself form? How does the hailstone form within the storm? What’s the impact on the things on the ground? A lot of that is very basic atmospheric science and requires research at the academic level. A lot of that is not going to be done at the commercial level,” he explains. “These are problems that are best solved at the university level because they have the intellect and — more importantly — they have the time to be able to cycle through these types of research projects.”

A distinctive feature of CIRCS is that commercial partners help shape the research agenda directly — eliminating the translation gap in which academic findings have to be repackaged before insurers can act on them. “It’s really nice for us to be able to give some insights and help push the research and the analysis in a direction that will make it directly applicable to what we need for our work,” says Suz Tolwinski-Ward, Head of Climate Statistics at Verisk.

This influence is visible in the projects approved for the first cycle. One will build a 10-year historical database of hail probability estimates across the US. Another combines coarse global climate models with high-resolution regional simulations to create roughly one million synthetic hailstorms — giving insurers and cat modelers a new tool for estimating how often severe hail events should be expected under past, present, and future climate conditions. A third simulates how hailstorms have historically struck populated areas, and how future climate change — combined with urban sprawl — could drive up damage and losses.

One of the perks of membership is that the commercial partners get priority access to each project’s deliverables. “Everything that’s developed within the Center is under a royalty-free IP license for all of the contributing members,” says Gensini. “It’s a unique deal for them. They’re kind of funding a start-up research center, and they do have intellectual access to all of the data and methods that come out.”

What the partners do with those outputs is up to them. Most expect to fold them into existing models and data processes to create more accurate predictions of insured hail losses in high-risk regions. In turn, this could help insurers to better price SCS risk — and identify areas they may want to steer clear from.

“As an insurance company, your whole business is taking risks. So what you want to do is understand how much risk to take in various areas, and where one risk might offset another risk,” says Bennett at Mercury Insurance.

With CIRCS, insurers and risk modelers may finally get the storm and hail data they've long lacked — improving these determinations and strengthening the sector's capacity as a whole to support communities recovering from these events.

Thanks for reading!

Louie Woodall
Editor