Overview
Last year brought record hail, significant tornadoes, and destructive straight-line winds to the U.S. and made it clear that severe convective storm (SCS) are no longer a “secondary” peril.
The 2026 Cotality Severe Convective Storm (SCS) Report reveals how a single hail event can produce losses that rival the insured financial fallout from a Category 4 hurricane. And they almost did in 2025. Our data reveals that a single hail event in the Dallas-Fort Worth area would have tripled had the storm drifted a mere 20 miles north.
Small distances have big impacts. That’s why insurance risk managers cannot rely on historical data alone. They need mathematical precision that includes future possibilities, not gut feel based only on past experience.
The anatomy of a modern probabilistic cat model
The year 1992 began the industry’s move toward probabilistic (stochastic) models, which could simulate thousands of potential events for any given region, and providing insight to underwriters on tail events that haven't yet happened but are physically possible. That was the year Hurricane Andrew blew by the industry’s worst-case scenarios, surpassing $15 billion in insured losses.
"Nothing in the historical record came close to predicting a Category 5 hurricane hitting Miami, or what the damages could equate to,” Cotality’s Chief Actuary Howard Kunst said. "Insurers weren’t prepared because their reliance on the historical record made it nearly impossible to predict the magnitude of loss that would result from a Category 5 hurricane hitting Southeast Florida. The increased exposure from the buildup of Miami had made it a whole different beast.”
The staggering losses from this storm was a pivotal shift in cat modeling. Today insurers have a different outlook. Cotality’s modern cat models allow you to see every policy in a portfolio, providing a comprehensive view of risk before an event occurs.
Our models are built upon four modules that translate potential catastrophic events into losses and the probability that they might occur. Within each model are hundreds of thousands of simulated, yet realistic events. This allows a user to calculate a potential dollar loss for every outcome — from the best-case minor event to the worst-case disaster.
The Hazard Module
The first module of any probabilistic cat model is the hazard module, which simulates the actual physical event. Using hurricanes as an example, this is the module that determines where a potential storm will form, travel, make landfall, and at what intensity. The module also determines frequency, defining how often a specific storm is expected to occur over a set period.
Frequency has become an essential consideration following the impacts of the 2017 and 2018 California wildfire seasons, which drove state regulatory bodies to become more deeply involved in regulating how insurers set rates for catastrophic risk.
The Exposure Module
You can’t have losses if there is nothing to be damaged. Our exposure module is where the user pinpoints the exact location and value of assets, whether they be homes, factories, or automobiles. It is here where building characteristics and reconstruction cost value (RCV) are entered. The more detail that can be entered the better, as the best results rely on the most precise and accurate exposure data. The 1994 Northridge earthquake, which devastated California communities, illustrated that fact, marking a shift toward models that included granular building intelligence.
The Vulnerability Module
The vulnerability module calculates the amount of physical damage. It considers the hazard — wind speed, water depth, ground shaking intensity — and the specific structural traits like roof materials or building age.
The Financial Module
The financial module translates physical destruction into dollars and cents. The financial module considers things like policy conditions, such as deductibles, limits, and exclusions, to quantify the insured losses.
Cotality’s models change constantly to consider not only evolving hazard behavior but also integrate new mitigation findings and updated building codes. These solutions are now used widely by insurers and reinsurers to navigate an increasingly complex landscape.
Speaking the language of risk in cat modeling
When you read Cotality’s 2026 Severe Convective Storm report, you’ll notice some terms that cat modelers, risk managers, and reinsurers use every day:
- Exceedance Probability (EP) is the probability that a certain level of loss will occur or be exceeded. For example, a model might show a 2% EP for a $100 million loss. This doesn’t mean the loss will be exactly $100 million; it means there is a 2% chance in any given year that losses will be $100 million or more.
- Return Period: This is the context for terms like “100-Year Storm.” It is another measure of probability; when something is deemed to be a 100-year event, there is a one-in-one hundred, or 1%, chance of that event occurring in any given year.
- The 50 vs. 500 Strategy: These are important data points for risk management. Insurers use the 50-year return period loss to stay profitable and a 500-year return period loss to remain solvent.
Cat modeling is the bedrock of modern risk management. With only 20 miles representing the difference between a manageable loss and a capital-depleting event, the importance of model literacy cannot be overstated.
Keep your mind open to what is possible in the ongoing evolution of these solutions.
