Assessing Age-Specific Contributions to Life Expectancy Gaps: The Proportional Difference Comparison Method

Antonino Polizzi, University of Oxford
Mathew Kiang, Stanford University
Monica Alexander, University of Toronto

A central goal of demographic analysis is to better understand life expectancy gaps between populations, such as by sex, socioeconomic status, or race/ethnicity. Because life expectancy increases more if the same proportional improvement occurs at younger vs. older ages, all else being equal, existing demographic decomposition methods may lead researchers to respectively overstate and understate the importance of mortality differences at younger and older ages. Here, we propose to compare (a) observed age-specific contributions to a given life expectancy gap with (b) hypothetical age-specific contributions that would be observed if the mortality rates between two populations differed by the same relative factor at all ages, but the life expectancy gap remained fixed. We apply the Proportional Difference Comparison Method (PDCM) to the Black-white life expectancy gap in the United States, showing that observed contributions of infant and young adult mortality exceed expected contributions to the gap by the same factor. In contrast, contributions of older ages are smaller than expected under the proportional scenario. We argue that PDCM can be used in comparative mortality analysis to better understand the biological and social factors shaping age-specific contributions to life expectancy gaps between populations.

Keywords: Mortality and Longevity, Decomposition analysis, Mathematical demography , Data and Methods

See extended abstract.