Incorporating Vitagions into Stochastic Longevity Models

Assessing mortality dynamics remains a central challenge in demographic, actuarial, and public health research, primarily due to the difficulty of producing reliable forecasts. This complexity is particularly highlighted during periods with large deviations from long-term trends, such as sudden mortality increases due to pandemics or decreases driven by medical and technological breakthroughs. In this context, we address the need for enhanced stochastic models that integrate sudden and significant mortality improvements, termed “vitagions”, into forward- looking forecasts. While foundational stochastic mortality models, such as those developed by Cairns et al. (2006b,a, 2009), provide a robust basis, they do not explicitly account for exogenous, forward-looking innovations. Building on this literature, we incorporate vitagions, defined as stochastic agents of mortality improvement associated with biomedical innovation and other external shocks (Woo (2014); Carannante et al. (2024)). Vitagions capture health-related advances, from disease prevention to advanced therapies, that can trigger persistent and one-sided mortality reductions. This paper makes one main contribution. We extend the Lee–Carter (LC) and Cairns- Blake-Dowd M6 (CBD M6) models by adding an exogenous innovation covariate that captures age-specific sensitivity to a detrended indicator of pharmaceutical innovation (constructed from FDA/EMA approvals), and we address identifiability through explicit normalisations and orthogonality constraints. An empirical analysis employs data from the U.S., France, and Italy, which show different historical mortality patterns. Scenario-based simulations deliver clearer period dynamics while preserving baseline age profiles.