Actuariat et Risque Contemporains
Time : 14h00 – 15h00
Date : 7th  February 2025

Salle des Chaires de l’Institut Louis Bachelier

Karim BARIGOU (Institut de Statistique, Biostatistique et Sciences Actuarielles, Université Catholique de Louvain) “Adapting mortality models to pandemic and climate shocks: A Bayesian vanishing jump and regime-switching framework”

Abstract: This presentation explores two complementary frameworks for advancing mortality modeling under extreme conditions. The first framework focuses on pandemic-related shocks and extends the Lee-Carter model to include vanishing jump effects. These effects capture the high initial impact of pandemics, such as COVID-19, followed by a gradual dissipation over time. Using a Bayesian approach, the model quantifies parameter uncertainty and outperforms existing approaches that assume transitory or persistent shock effects. The second framework addresses both temperature- and epidemiological-related shocks through a granular regime-switching model with three states: (1) baseline seasonal mortality patterns, (2) environmental shocks (e.g., heatwaves), and (3) epidemiological shocks (e.g., influenza). Transition probabilities are modeled using covariate-dependent multinomial logit functions, and the framework is calibrated on high-resolution weekly mortality data from NUTS 2 regions in France. These approaches provide actionable insights for forecasting mortality under extreme events, offering practical applications for insurers, healthcare providers, and policymakers.

Organisateurs : Hillairet Caroline, Olivier Lopez

Lieu :

Salle des Chaires de l’Institut Louis Bachelier

4e étage, Palais Brongniart

16 place de la Bourse

Paris 2eme.