LONGFOR · The future of longevity: A theoretical framework to inform mortality forecasting

Current mortality forecasting methods have been shown to systematically underpredict future longevity, and their methodological frameworks are based on time series extrapolation rather than theory or structural models. Given our extensive knowledge of the factors that have shaped mortality trends, a crucial question is whether considering additional traits of the populations analysed can improve prediction accuracy compared to the state of the art.The goal of LONGFOR is to develop innovative models of past and future mortality for several populations that are informed by the theoretical framework of the health transition theory. To achieve this, LONGFOR will focus on three primary research objectives: (A) to determine whether the mortality transitions of European populations stratified by sex since the 1950s have shared common patterns in terms of their mortality determinants, (B) to develop a novel theory-driven model to describe and forecast age-specific mortality by sex of multiple populations based on their mortality determinants, and (C) to extend this multi-population model to cause-specific mortality forecasting by age and sex.LONGFOR will achieve these objectives by integrating advanced statistical methods with publicly available data on mortality, fertility, population, causes of death, socio-economic, behavioural and environmental indicators. This ambitious but feasible project has the potential to make ground-breaking contributions to both mortality theory and the practice of mortality forecasting. Specifically, LONGFOR findings will have significant societal and scientific impacts by: (i) providing empirical evidence on current mortality theories, (ii) offering new insights for theoretical and modelling advancements, (iii) contributing innovative methods to forecast mortality, and (iv) improving current forecasts of both all-cause and cause-specific mortality.