Volume 37 - Article 17 | Pages 527–566 Author has provided data and code for replicating results

Coherent forecasts of mortality with compositional data analysis

By Marie-Pier Bergeron-Boucher, Vladimir Canudas-Romo, Jim Oeppen, James W. Vaupel

Print this page  Facebook  Twitter

 

 
Date received:26 Oct 2016
Date published:29 Aug 2017
Word count:5203
Keywords:coherent mortality forecasting models, compositional data, Lee-Carter model, life expectancy, Li-Lee, mortality forecasts
DOI:10.4054/DemRes.2017.37.17
Additional files:readme.37-17 (text file, 1 kB)
 demographic-research.37-17 (zip file, 5 kB)
 

Abstract

Background: Mortality trends for subpopulations, e.g., countries in a region or provinces in a country, tend to change similarly over time. However, when forecasting subpopulations independently, the forecast mortality trends often diverge. These divergent trends emerge from an inability of different forecast models to offer population-specific forecasts that are consistent with one another. Nondivergent forecasts between similar populations are often referred to as "coherent."

Methods: We propose a new forecasting method that addresses the coherence problem for subpopulations, based on Compositional Data Analysis (CoDa) of the life table distribution of deaths. We adapt existing coherent and noncoherent forecasting models to CoDa and compare their results.

Results: We apply our coherent method to the female mortality of 15 Western European countries and show that our proposed strategy would have improved the forecast accuracy for many of the selected countries. The results also show that the CoDa adaptation of commonly used models allows the rates of mortality improvements (RMIs) to change over time.

Contribution: This study opens a discussion about the use of age-specific mortality indicators other than death rates to forecast mortality. The results show that the use of life table deaths and CoDa leads to less biased forecasts than more commonly used forecasting models based on the extrapolation of death rates. To the authors’ knowledge, the present study is the first attempt to forecast coherently the distribution of deaths of many populations.

Author's Affiliation

Marie-Pier Bergeron-Boucher - Syddansk Universitet, Denmark [Email]
Vladimir Canudas-Romo - Australian National University, Australia [Email]
Jim Oeppen - Max Planck Odense Center on the Biodemography of Aging, Denmark [Email]
James W. Vaupel - Syddansk Universitet, Denmark [Email]

Other articles by the same author/authors in Demographic Research

» In Memoriam: Professor Jan M. Hoem
Volume 36 - Article 24

» Decomposing changes in life expectancy: Compression versus shifting mortality
Volume 33 - Article 14

» Demographic characteristics of Sardinian centenarian genealogies: Preliminary results of the AKeA2 study
Volume 32 - Article 37

» The Gompertz force of mortality in terms of the modal age at death
Volume 32 - Article 36

» Maternal longevity is associated with lower infant mortality
Volume 31 - Article 42

» Unobserved population heterogeneity: A review of formal relationships
Volume 31 - Article 22

» Cause-specific measures of life years lost
Volume 29 - Article 41

» The difference between alternative averages
Volume 27 - Article 15

» The crossover between life expectancies at birth and at age one: The imbalance in the life table
Volume 24 - Article 4

» Attrition in heterogeneous cohorts
Volume 23 - Article 26

» Senescence vs. sustenance: Evolutionary-demographic models of aging
Volume 23 - Article 23

» Total daily change with age equals average lifetime change
Volume 22 - Article 36

» No consistent effects of prenatal or neonatal exposure to Spanish flu on late-life mortality in 24 developed countries
Volume 22 - Article 20

» Survival as a Function of Life Expectancy
Volume 21 - Article 29

» The age separating early deaths from late deaths
Volume 20 - Article 29

» Life lived and left: Carey’s equality
Volume 20 - Article 3

» Formal Relationships: Introduction and Orientation
Volume 20 - Article 1

» An integrated approach to cause-of-death analysis: cause-deleted life tables and decompositions of life expectancy
Volume 19 - Article 35

» The modal age at death and the shifting mortality hypothesis
Volume 19 - Article 30

» The relative tail of longevity and the mean remaining lifetime
Volume 14 - Article 7

» Lifesaving, lifetimes and lifetables
Volume 13 - Article 24

» Changing mortality and average cohort life expectancy
Volume 13 - Article 5

» Age-specific contributions to changes in the period and cohort life expectancy
Volume 13 - Article 3

» Oldest Old Mortality in China
Volume 8 - Article 7

» Life Expectancy at Current Rates vs. Current Conditions: A Reflexion Stimulated by Bongaarts and Feeney’s "How Long Do We Live?"
Volume 7 - Article 8

» Decomposing demographic change into direct vs. compositional components
Volume 7 - Article 1

» Dr. Väinö Kannisto: A Reflexion
Volume 6 - Article 5

Most recent similar articles in Demographic Research

» Point and interval forecasts of mortality rates and life expectancy: A comparison of ten principal component methods
Volume 25 - Article 5    | Keywords: Lee-Carter model, life expectancy

» High life expectancy and reversed socioeconomic gradients of elderly people in Mexico and Costa Rica
Volume 38 - Article 3    | Keywords: life expectancy

» Summertime, and the livin’ is easy: Winter and summer pseudoseasonal life expectancy in the United States
Volume 37 - Article 45    | Keywords: life expectancy

» Macroeconomic conditions and population health in Iceland
Volume 37 - Article 25    | Keywords: life expectancy

» The population-level impact of public-sector antiretroviral therapy rollout on adult mortality in rural Malawi
Volume 36 - Article 37    | Keywords: life expectancy