Stress testing insurance market stability under climate risk

Simona Meiler, Steven I Jackson, Kerry Emanuel, Noah S Diffenbaugh, Jack W. Baker

Extreme weather events exert increasing pressure on communities in hazard-prone areas and on the systems designed to protect them. Insurance serves as a risk-transfer mechanism, providing financial security for homeowners and supporting community resilience. Yet, behind this first layer of protection lies a complex web of reinsurers, capital markets, and public institutions that absorb and redistribute disaster risk. Intensifying climate hazards, continued coastal development, and evolving market dynamics threaten the stability of this network.

Here, we develop a risk-propagation model to assess whether single or sequential tropical cyclones striking Florida could generate systemic financial stress across the property-insurance system. The model links physics-based, probabilistic simulations of tropical cyclone wind and flood losses with data on the Florida residential insurance market, its backstop mechanisms, and regulatory frameworks. By tracing how losses cascade as capital constraints and contractual thresholds bind, we evaluate systemic risk under present-day conditions, future climate scenarios, and alternative market and adaptation configurations within the same quantitative architecture.

We estimate an annual 4\% probability that total public burden exceeds 1\% of Florida’s GDP, indicating that extreme tropical cyclone seasons can overwhelm private market capacity and shift costs to public institutions. Vulnerabilities differ across institutional layers, with post-insolvency mechanisms, residual markets, and the NFIP exhibiting greater sensitivity to extreme and sequential events than formal reinsurance limits. Although we focus on Florida in this proof-of-concept study, this approach provides a transferable template for quantitatively stress-testing insurance systems and evaluating how climate change, market dynamics, and adaptation strategies reshape systemic disaster risk.