Dr. Ayse Arik, University of New South Wales, Sydney
Title: Modelling Post-pandemic Mortality: Structural trend changes or extreme shocks?
Date: Friday, October 30th, 2026
Time: 1:30PM (PDT)
Location: ASB 10900
Abstract: The projection of mortality rates is central to the assessment of liabilities in life and pension insurance portfolios. COVID-19 has made this task more demanding, raising fundamental questions about whether standard projection models remain fit for purpose in a world of structural trend changes and extreme mortality shocks.
We propose a simulation framework extending the age-period-cohort model along two fronts: explicit modelling of structural trend changes, calibrated from data on 48 populations across 24 countries from the Human Mortality Database; and a family of shock distributions of increasing tail heaviness 鈥� Gaussian, Normal-Inverse Gaussian, and Peaks-Over-Threshold with a Generalised Pareto tail. This framework is transparent and practically oriented, and is implemented in the freely available R package trendMoMo.
Our results show that long-term projection uncertainty is primarily driven by uncertainty in future structural trend changes, whereas uncertainty arising from period shocks has only a marginal impact on long horizons and is mainly relevant over shorter projection periods. These findings have important implications for risk management: unexpected mortality shocks are particularly relevant for short-term solvency assessments, while uncertainty in the underlying evolution of mortality trends represents a greater concern for insurance pricing, reserving, and long-term longevity risk assessment.