Swedish Mortality Does Not Decelerate in Recent Cohorts

Roland Rau , University of Rostock & Max Planck Institute for Demographic Research
Marcus Ebeling, Karolinska Institutet
Torsten Sauer, University of Rostock
Karin Modig, Karolinska Institutet
Anders Ahlbom, Karolinska Institutet

Background: Mortality at the highest ages tends to level off, a phenomenon typically described as mortality deceleration. The standard explanation is mortality selection. There are claims in the literature, however, that the observed mortality deceleration is an artifact. Objective: By using data of the highest quality, which do not suffer from any of the problems claimed, we wanted to show that mortality does, indeed, decelerate. Data & Method: We used Swedish cohort data for birth years 1890 through 1917. All required data were available by exact date to be able to estimate exact person-years of exposure. We plotted the resulting death rates and fitted Gompertz- and Gamma-Gompertz-models to the observed data. Results: While we found mortality deceleration for older birth cohorts, none of the younger birth cohorts showed any signs of mortality deceleration. This finding contradicts our expectations completely and requires further investigation.

See extended abstract

 Presented in Session 17. Oldest Old