Glossary

Catastrophe model

Software that simulates thousands of hurricanes, earthquakes and other disasters against a portfolio of locations to estimate insured losses and odds.

Also called: cat model, nat cat model, catastrophe modeling, cat modeling

A catastrophe model estimates the losses a portfolio of insured property might suffer from natural disasters. Vendors such as Verisk and Moody's RMS build models in three parts: a hazard module that generates a stochastic set of simulated events (hurricanes, earthquakes, wildfires, floods, severe convective storms), a vulnerability module that estimates damage by construction type, age and occupancy, and a financial module that applies deductibles, limits and reinsurance to produce insured loss.

The outputs are probabilities, not forecasts. Average annual loss is the expected yearly cost; the exceedance probability curve shows the chance that losses reach a given amount in a year; a 1-in-100-year loss is the point with a 1% annual chance of being exceeded. Insurers use them to price, manage concentrations in one region, buy reinsurance and satisfy rating agencies. Underwriters feed them the occupancy, construction, year built and values from your statement of values.

For a business buyer, the model is why a coastal or wildfire-zone location can see pricing or terms change when the vendor updates its view of risk, even with no losses of your own. Model output is only as good as the data: wrong construction class or understated values distorts the result and the price. The guide on catastrophe models: how insurers price hurricanes, quakes and fires explains how to prepare data, and probable maximum loss describes the related measure.

Last reviewed October 10, 2026

In the index now

Related terms

Related guides