Catastrophe risk models are increasingly incorporating current climate and environmental conditions rather than relying primarily on long-term historical averages, creating new opportunities and challenges for insurers and reinsurers, according to Howden Re.
Updates from Verisk and forthcoming developments from Moody’s RMS are changing how the industry estimates potential losses from hurricanes, wildfires and severe convective storms.
The shift allows insurers to assess catastrophe exposure under different climate risk conditions and develop more tailored views of portfolio risk. However, Howden Re cautions that greater modelling sophistication does not necessarily produce more reliable loss estimates.
Key Highlights
- Catastrophe modelling is shifting toward current and near-term climate conditions, with Verisk updating its hurricane, wildfire and severe convective storm models.
- Moody’s RMS plans to introduce similar climate-sensitive capabilities in its next-generation hurricane model in 2027, incorporating ENSO and North Atlantic Oscillation indicators.
- Howden Re warns that changes in modelled catastrophe losses do not necessarily indicate equivalent changes in underlying risk, making independent analysis essential.
Global economic losses from natural disasters in the 9 months of 2026 were estimated at approximately $204 bn, which was 31% below the long-term average since 2000 ($296 bn) and 25% below the median of the same period ($271 bn). It was also the lowest Q1-Q3 economic loss since 2015.
Catastrophe models increasingly reflect present-day climate conditions
Traditional catastrophe models use historical events, scientific research and simulated scenarios to estimate the frequency and severity of natural disasters.
Newer models increasingly incorporate current climate conditions and environmental changes that may influence future catastrophe activity.
Verisk has introduced a near-present climate approach in its updated hurricane model, following similar changes to its wildfire and severe convective storm models.
The methodology gives greater consideration to recent and near-term climate conditions rather than depending exclusively on historical averages.
Moody’s RMS is expected to introduce a comparable approach in its next-generation high-definition hurricane model, scheduled for release in 2027.
Anna Pergerson, Managing Director and Head of Catastrophe R&D at Howden Re, said different model releases can produce different interpretations of historical data. Because modelling updates are introduced separately across different perils, insurers must assess portfolios using models built on potentially different assumptions.
Understanding the causes of changes in modelled losses is therefore becoming increasingly important.
Verisk and Moody’s RMS incorporate ENSO and ocean temperature indicators
New catastrophe models provide greater visibility into the climate variables influencing simulated events. Verisk’s updated hurricane model allows simulated years to be associated with specific phases of the El Niño-Southern Oscillation (ENSO) and Atlantic sea surface temperatures.
Insurers can analyze subsets of simulated events associated with particular climate conditions, allowing them to examine how hurricane losses might differ under alternative scenarios.
Moody’s RMS plans to incorporate ENSO into its forthcoming hurricane model, alongside the North Atlantic Oscillation (NAO), which can influence storm tracks. These capabilities allow insurers and reinsurers to test portfolio sensitivity to different atmospheric and ocean conditions.
The ability to examine a broader range of outcomes can improve decisions about underwriting, portfolio management and reinsurance purchasing.
Sid Sasanian, Managing Director at Howden Re
However, the usefulness of these analyses depends on the quality of the underlying assumptions and the interpretation of model results.
Howden Re uses proprietary analytical tools to supplement commercial catastrophe models. Its Blaze Index assesses present-day wildfire risk using current information about vegetation, fuel conditions, climate and the interaction between communities and surrounding environments.
Unlike approaches based primarily on historical wildfire events, the index updates its assessment as relevant environmental and exposure data change.
More detailed catastrophe models do not eliminate uncertainty
Howden Re emphasizes that increasing model complexity does not automatically reduce uncertainty. This is particularly relevant for geographically concentrated insurance portfolios.
When exposure is concentrated in a limited area, the number of simulated catastrophe events affecting that location may be insufficient to support reliable conclusions.
Filtering model results by specific climate conditions can further reduce the number of relevant simulated events.
Consequently, more detailed analysis may provide additional insight without necessarily increasing statistical confidence.
Insurers also need to distinguish between changes in actual catastrophe exposure and changes caused by revised modelling methods.
For example, an increase in expected losses following a model update may reflect new climate assumptions, revised vulnerability estimates or changes in how simulated events are represented. It does not necessarily mean the underlying risk has increased by the same amount. Likewise, unchanged modelled losses do not establish that exposure has remained stable.
What changes mean for insurance underwriting and reinsurance buying
The latest catastrophe modelling developments have implications for underwriting, capital management and reinsurance purchasing.
Climate-sensitive models allow insurers to examine potential loss outcomes under different environmental conditions and assess whether existing portfolio assumptions remain appropriate.
They can also support decisions about geographical exposure, risk selection and the structure of catastrophe reinsurance programs. However, model updates can change expected losses even when an insurer’s underlying portfolio has not materially changed.
This creates a need to reconcile differences between model versions and assess the financial consequences before adjusting underwriting or reinsurance strategies.
Reinsurers recommends combining commercial catastrophe models with independent analysis and specialist expertise. No single model can capture every dimension of catastrophe risk.
As insurers adopt more sophisticated modelling tools, understanding the assumptions behind their results will remain as important as the loss estimates themselves.
………………..
EXPERTS: Anna Pergerson – Managing Director and Head of Catastrophe R&D at Howden Re, Sid Sasanian – Managing Director at Howden Re
by Yana Keller – Re/Insurance Editor at Beinsure
