A generalized two‐factor square‐root framework for modeling occurrences of natural catastrophes

This work aims to forecast (over 1, 5, and 15 years) the extremes, expected value, volatility of natural disasters occurrences. To achieve this objective, we adopt a generalized two-factor square-root model linking together occurrences through stochastic correlation (Brownian motion). We use Pareto distribution (GPD) maximum number as measure value at risk (VaR). The results are checked in terms accuracy, compared versus some baseline models (i.e., Poisson process extreme model) backtested.