The Black–White Mortality Crossover: Evidence from Linked U.S. Administrative Mortality Records

• In Event: Morbidity and Mortality: Open Call Paper Session

Abstract

First documented in 1930, the Black–White mortality crossover remains one of the most enduring puzzles in Demography. Black Americans experience higher age-specific mortality rates than White Americans throughout most of the life course, yet this pattern reverses at advanced ages. A leading explanation is that there is no crossover at all. Rather, differential age misreporting or other data errors create a spurious crossover. We use a large-scale administrative cohort dataset of American men (N = 2.3 million deaths) to assess whether the Black–White mortality crossover is real or an artifact of data error. We find a mortality crossover for the male birth cohorts of 1890–1905 at age 86. At age 75, Black men have nearly a 10% higher probability of dying than White men, but by age 95, the probability of dying is 8% higher for White men. The quality of our mortality data, paired with a series of sensitivity analyses, provides new evidence that the observed crossover is not an artifact of age misreporting. Our findings address a long-standing empirical debate and emphasize the need for further investigation into alternative explanations of the crossover, including selective survival mechanisms and reversals of group-specific disadvantage at advanced ages.

Author

• Casey Breen, The University of Texas at Austin