Dissertation

The causes and the consequences of growing life expectancy shortfalls

Abstract

Life expectancy stagnation the lack of improvement in life expectancy over timeand life expectancy divergence the widening of an existing life expectancy gap are both warning signs that population health is not developing as well as it could. Most demographic research has used a stagnation perspective to understand recent life expectancy trends in high-income countries. This dissertation argues that a within-country approach risks overlooking critical mortality dynamics that are unique to the population experiencing life expectancy stagnation. It proposes to complement within-country analysis of life expectancy stagnation with betweencountry analysis of life expectancy divergence. In three empirical studies, this dissertation identifies the causes of recent life expectancy trends in the United States of America (USA) and the United Kingdom (UK) relative to other high-income countries before, during, and after the COVID-19 pandemic, and quantifies the consequences of this life expectancy divergence for population composition. Using demographic decomposition, Chapter 2 identifies the age groups and causes of death that contributed to the decreasing life expectancy advantage or increasing life expectancy disadvantage of England and Wales relative to other European countries in the period 2010 2019. Chapter 3 identifies the age groups and causes of death that contributed to life expectancy divergence between the USA and the UK and betterperforming peer countries during the COVID-19 pandemic. Chapter 4 considers the consequences of the growing life expectancy shortfall of the USA. Using counterfactual methods, it estimates the number of children that were never born because the USA consistently experienced higher mortality than its peers. Across the three chapters, the proposed divergenceapproach points to critical population dynamics in the USA and the UK that had previously gone unnoticed with a stagnation approach. Thus, this dissertation argues that the joint application of the two approaches should become more common practice in demographic analysis.