Author(s): Hauser, Andrea ; Rosa, Carlos ; Esteves, Rui Esteves ; Moura, Alexandra ; Oliveira, Carlos
Date: 2021
Persistent ID: http://hdl.handle.net/10400.5/22775
Origin: Repositório da Universidade de Lisboa
Subject(s): Risk; Hurricanes; Property Insurance; Regression Models
A complete model to analyse and predict future losses in the property portfolio of an insurance company due to hurricanes is proposed. A novel statistical model, in which weather data is not required, is considered. Climate data may not be reliable, or may be difficult to deal with or to obtain, hence we reconstruct the storm behaviour through the registered claims and respective losses. The model is calibrated using the loss data of the property portfolio of the insurance company Fidelidade, from hurricane Leslie, which hit the center of continental Portugal in October 2018. Several scenarios are simulated and risk maps are built. The simulated scenarios can be used to compute risk premiums per risk class in the portfolio. These can be used to adjust the policy premiums accounting for a storm risk. The risk map of the company also depends on its portfolio, namely its exposure, providing a hurricane risk management tool for the insurance company.
