Background: Blacks have lower life expectancy than whites in the United States. That disparity could be due to racial differences in the causes of death, with blacks being more likely to die of causes that affect the young, or it could be due to differences in the average ages of blacks and whites who die of the same cause. Prior studies fail to distinguish these two possibilities.

Objective: In this study we determine how much of the 2000-10 reduction in the racial gap in life expectancy resulted from narrowing differences in the cause-specific mean age at death for blacks and whites, as opposed to changing cause-specific probabilities for blacks and whites.

Methods: We introduce a method for separating the difference-in-probabilities and difference-in-age components of group disparities in life expectancy.

Results: Based on the new method, we find that 60% of the decline in the racial gap in life expectancy from 2000 to 2010 was attributable to reduction in the age component, largely because of declining differences in the age at which blacks and whites die of chronic diseases.

Conclusions: Our findings shed light on the sources of the declining racial gap in life expectancy in the United States, and help to identify where advances need to be made to achieve the goal of eliminating racial disparities in life expectancy.