Volume 33 - Article 21 | Pages 589–610

Taylor's power law in human mortality

By Christina Bohk, Roland Rau, Joel E. Cohen

Print this page  Facebook  Twitter


Date received:10 Nov 2014
Date published:17 Sep 2015
Word count:2500
Keywords:death rates, rates of mortality improvement, Taylor’s law, variance to mean relationship
Additional files:33-21 high resolution figures for download (zip file, 9 MB)


Background: Taylor's law (TL) typically describes a linear relationship between the logarithm of the variance and the logarithm of the mean of population densities. It has been verified for many non-human species in ecology, and recently, for Norway’s human population. In this article, we test TL for human mortality.

Methods: We use death counts and exposures by single age (0 to 100) and calendar year (1960 to 2009) for countries of the Human Mortality Database to compute death rates as well as their rates of change in time. For both mortality measures, we test temporal forms of TL: In cross-age-scenarios, we analyze temporal variance to mean relationships at different ages in a certain country, and in cross-country-scenarios, we analyze temporal variance to mean relationships in different countries at a certain age.

Results: The results reveal almost log-linear variance to mean relationships in both scenarios; exceptions are the cross-country-scenarios for the death rates, which appear to be clustered together, due to similar mortality levels among the countries.

Conclusions: TL appears to describe a regular pattern in human mortality. We suggest that it might be used (1) in mortality forecasting (to evaluate the quality of forecasts and to justify linear mortality assumptions) and (2) to reveal minimum mortality at some ages.

Author's Affiliation

Christina Bohk - Max-Planck-Institut für Demografische Forschung, Germany [Email]
Roland Rau - Universität Rostock, Germany [Email]
Joel E. Cohen - Rockefeller University, United States of America [Email]

Other articles by the same author/authors in Demographic Research

» Measuring the concentration of urban population in the negative exponential model using the Lorenz curve, Gini coefficient, Hoover dissimilarity index, and relative entropy
Volume 44 - Article 49

» Gompertz, Makeham, and Siler models explain Taylor's law in human mortality data
Volume 38 - Article 29

» Is the fraction of people ever born who are currently alive rising or falling?
Volume 30 - Article 56

» Minor gradient in mortality by education at the highest ages: An application of the Extinct-Cohort method
Volume 29 - Article 19

» Life expectancy: Lower and upper bounds from surviving fractions and remaining life expectancy
Volume 24 - Article 11

» Life expectancy is the death-weighted average of the reciprocal of the survival-specific force of mortality
Volume 22 - Article 5

» Constant global population with demographic heterogeneity
Volume 18 - Article 14

» Seasonal mortality in Denmark: the role of sex and age
Volume 9 - Article 9

Similar article in Demographic Research

» Leveraging deep neural networks to estimate age-specific mortality from life expectancy at birth
Volume 47 - Article 8    | Keywords: death rates