The Contribution of Cohort and Period Effects to Historical Mortality Decline in Six Developed Countries

• In Event: 539 - Regular Session. Methodological Challenges of Research on Mortality and Morbidity

Abstract

Whether cohort or period forces contribute to historical mortality declines in developed countries is an unresolved question. In this study, we view cohort and period as latent constructs that influence mortality risks over the life course within a cohort, and influence mortality risks across all age groups within a period, respectively. Guided by this conceptualization, we use factor analysis to extract the eigenvalue of the cohort (or period) construct from cohort (or period) age-specific mortality rates. We examine mortality rates in six developed countries, and find that the strength of the cohort construct took off in the mid-19th century birth cohort and increased or remained stable in most combinations of country and gender. We also find that the strength of the period construct generally increased before the 1940s, especially among women; declined between the 1940s and 1960s; and rebounded after the 1970s. These findings suggest that the mortality decline between the mid-19th century and early 21st century is a result of both period and cohort effects, except during the period spanning the 1940s through the 1960s, when cohort effects predominated as the source of mortality decline. Therefore, disregarding either period or cohort effects may lead to underestimating the extent of mortality decline in the past and in the future.

Authors

• Hui Zheng, The Ohio State University

• Scott M. Lynch, Duke University