Introduction:
The increasing frequency and severity of extreme precipitation events pose significant challenges for the insurance industry in quantifying and managing climate-related economic losses. This study develops a statistical–actuarial framework to estimate the distribution of economic losses conditional on extreme precipitation levels. The approach integrates heavy-tailed marginal modeling with copula-based dependence structures, allowing the joint behavior of meteorological extremes and disaster losses to be analyzed within a coherent probabilistic setting.

Methods:
To model the marginal behavior of precipitation and loss variables, we assess a range of right‑skewed candidate distributions, including those with heavy fat tails. Dependence between precipitation and losses is captured using Gaussian, t-Student, and Gumbel copulas. Conditional loss quantiles, including Value at Risk (VaR), are estimated under different precipitation scenarios. Weekly maximum precipitation data from Spain (2018–2025) and economic loss records from the Emergency Events Database (EM-DAT) for selected European countries and the NOAA Storm Events Database for the United States are used for empirical illustration.

Results:
The results show that specifications incorporating upper-tail dependence generate substantially higher conditional Value at Risk estimates than symmetric Gaussian dependence, particularly in the United States, where extreme losses dominate the distribution.

Conclusions:
The proposed framework provides a flexible and robust tool for climate risk assessment, supporting insurers and policymakers in capital allocation, pricing, and disaster risk management under extreme weather conditions.