Catastrophe Modeling Service Bureaus for Reinsurance
Catastrophe modeling sits at the core of modern reinsurance underwriting. Service bureaus — vendors that develop, license, and run probabilistic catastrophe models — provide the quantitative foundation for pricing property catastrophe treaties, managing aggregate exposures, and meeting regulatory capital requirements.
The Big Three and Beyond
For decades, three vendors dominated the market: AIR Worldwide (now Verisk Extreme Event Solutions), RMS (now Moody’s RMS), and CoreLogic. Together they account for the vast majority of licensed cat model usage across primary insurers and reinsurers globally. Each offers 100+ country models spanning earthquake, tropical cyclone, flood, severe convective storm, wildfire, and winter storm perils.
However, the landscape has expanded significantly. Karen Clark & Company (KCC), founded by the creator of the original catastrophe model, offers RiskInsight — an open loss modeling platform covering major perils in 50+ countries. The Oasis Loss Modelling Framework, a not-for-profit backed by nearly 40 major insurers and reinsurers, hosts 90+ models from 18+ suppliers on a fully open-source simulation engine, lowering barriers for new entrants.
Emerging and Specialist Vendors
The next generation of cat modelers targets specific gaps left by the incumbents:
- Flood Specialists
- JBA Risk Management, Fathom, and KatRisk provide high-resolution inland flood and storm surge models that often outperform legacy vendor outputs in granularity.
- Climate-Forward Modelers
- Reask uses AI-driven probabilistic hazard mapping for tropical cyclones. ZestyAI delivers parcel-level risk scores using computer vision and machine learning.
- Regional Specialists
- COMBUS models Australian perils; ERN/RED covers Latin America; CatRisk focuses on European windstorm and earthquake; Applied Research Associates (ARA) provides U.S. hurricane and tornado models used by FEMA.
How Reinsurers Use Cat Model Outputs
Reinsurance underwriters and actuaries rely on cat model outputs at multiple stages:
| Stage | Application |
|---|---|
| Treaty Pricing | Exceedance probability (EP) curves drive occurrence and aggregate loss estimates for XoL and quota share structures |
| Portfolio Management | Probable Maximum Loss (PML) and Tail Value at Risk (TVaR) metrics inform risk appetite and accumulation limits |
| Capital Modeling | Stochastic event sets feed internal capital models and rating agency assessments (AM Best BCAR, S&P) |
| ILS Structuring | Cat bond sponsors require independent model runs — Verisk alone has modeled over 0 billion in catastrophe bonds |
Selecting a Cat Modeling Vendor
Key evaluation criteria for reinsurance buyers include:
- Peril and territory coverage — does the vendor cover all perils in your book?
- Model transparency — are hazard, vulnerability, and financial modules individually accessible and auditable?
- Regulatory acceptance — is the model approved or accepted by Solvency II internal model supervisors, APRA, or state DOIs?
- Integration — does the output feed into your reinsurance pricing platform (e.g., Sequel Impact, Aon PathWise, Guy Carpenter GC AdvantagePoint)?
- Update cadence — how often are event sets and vulnerability functions refreshed?