Efforts to understand the dramatic declines in mortality over the past century have focused on life expectancy. However, understanding changes in disparity in age of death is important to understanding mechanisms of mortality improvement and devising policy to promote health equity. We derive a novel decomposition of variance in age of death, a measure of inequality, and apply it to cause-specific contributions to the change in variance among the G7 countries (Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States) from 1950 to 2010. We find that the causes of death that contributed most to declines in the variance are different from those that contributed most to increase in life expectancy; in particular, they affect mortality at younger ages. We also find that, for two leading causes of death [cancers and cardiovascular disease (CVD)], there are no consistent relationships between changes in life expectancy and variance either within countries over time or between countries. These results show that promoting health at younger ages is critical for health equity and that policies to control cancer and CVD may have differing implications for equity.health equity | age of death variance | life expectancy | inequality | mortality T he 20th century was an epoch of unprecedented declines in mortality rates in most countries. Causes of death shifted from infections, which primarily killed children, to cardiovascular disease (CVD) and cancer, which primarily killed adults, in the epidemiological transition (1, 2). Understanding mortality change largely focused on understanding changes in the average age of death, the life expectancy (e 0 ). In recent years, attention has shifted to disparities in mortality, and health equity is now recognized as an important policy goal. Here, we develop and use a new method to show how changes in inequality, as measured by the variance V, are driven by changes in mortality caused by specific causes, accounting for the interdependence of changes in the different causes of death. We find differences between the magnitudes of cause-specific contributions to e 0 and V and unexpectedly inconsistent relationships between the two over time and across countries, with important scientific and policy implications.Work to date on understanding change in e 0 has found that, among wealthy nations, much of the gain in the second half of the 20th century was because of declines in CVD mortality (3-5). The male-female difference in e 0 , which widened and then, narrowed over the course of the century, has also been driven by CVD mortality as a consequence of smoking behavior (6-8). Disparities across countries in e 0 have been studied most notably by the Global Burden of Disease Project, which has noted gains to life expectancy across countries and declines in global disparity for certain age-specific mortality rates (9, 10). Related work has shown how changes in different causes of death would have different effects on the health of global poor vs. rich (11).Analyses of al...