Integrated Frequency and Severity Modeling for Extreme Losses
Variance · 20.11.2026 · 19
DOI: 10.66573/001c.172255 ↗Kurz gefasst
Fachartikel in der Zeitschrift Variance, online erschienen am 20.11.2026. Verfasst von Haibo Liu und 2 weiteren Autor:innen. Die Arbeit nennt 19 Literaturangaben.
Abstract (Originalsprache)
The frequency and severity of climate-related losses have increased in recent decades, exhibiting clustering, time-varying risk, and frequency–severity dependence that pose challenges to traditional aggregate loss models. This study employs a self-exciting peaks-over-threshold model with self-excitation in both the occurrence intensity and the severity distribution, allowing for clustering, nonstationarity, seasonality, and formal testing of self-excitation effects. Using California wildfire loss data, formal statistical tests and short-horizon loss prediction are applied to show that this model provides the best fit relative to baseline models without self-excitation. This work was supported by a 2024 individual research grant from the CAS and the Society of Actuaries (SOA).
Zitieren
APA
Liu, H., Tong, Z., & Zhang, Y. (2026). Integrated Frequency and Severity Modeling for Extreme Losses. Variance, 19. https://doi.org/10.66573/001c.172255
Vancouver
Liu H, Tong Z, Zhang Y. Integrated Frequency and Severity Modeling for Extreme Losses. Variance. 2026;19. doi:10.66573/001c.172255
BibTeX
@article{Liu2026,
author = {Liu, Haibo and Tong, Zhiwei and Zhang, Yujin},
title = {Integrated Frequency and Severity Modeling for Extreme Losses},
journal = {Variance},
year = {2026},
volume = {19},
doi = {10.66573/001c.172255}
}
EUVIMED