Patterns of Mortality Improvement over Age and Time: Estimation, Presentation and Implications

Kirill F. Andreev, Queen's University
James W. Vaupel, Max Planck Institute for Demographic Research

Changes in mortality over time are usually analyzed by exploring trends in death rates or by calculating rates of mortality improvement. Typically rates of mortality improvement are computed over extended periods of time and over broad age groups as dictated by stability of estimates. However this approach might conceal important details of temporal mortality dynamics potentially hindering development of adequate mortality models for exploration, comparisons or forecasting mortality trends. In this work we estimated surfaces of mortality improvement for several developed countries by a smoothing tensor product spline. The results suggest that mortality did not decline uniform over time across the countries. Moreover strong cohort-like patterns have been detected for several populations. These findings have a potential importance for building better models of mortality forecasts. The origin of such cohort-like patterns is still obscure; further work is required to provide possible explanations for the observed patterns.