The U.S. Mortality Crisis as a Preston Curve Reversal

U.S. life expectancy stagnated and declined in the 2010s despite continued growth in real per capita income. We use Preston curves to characterize this pattern as a change in the relationship between income and longevity. Using state-level data from 1980 to 2019 and county-level data from 2000 to 2019, we estimate population-weighted Preston curves relating life expectancy to logged real per capita income. From 1980 to 2010, U.S. states followed the classic Preston curve pattern: rising income was accompanied by rising life expectancy. Counties followed the same pattern from 2000 to 2010. From 2010 to 2019, however, states and continued to become richer while life expectancy stagnated or declined. The curves shifted right without shifting up and became steeper, indicating decoupling and divergence: increases in aggregate resources over time no longer produced broad longevity gains, and, in any given year, inequality in life expectancy by income grew. These patterns are robust to alternative temporal anchors around the Great Recession and to substituting education for income. They also appear across sex and racial groups. County-level decompositions are broadly consistent with arguments that longevity has fallen due to widely shared exposure to social deterioration, which may account for the Preston curve reversal. Collectively, we show that the recent U.S. mortality crisis reveals a weakening -- and growing inequality -- in the conversion of aggregate resources into longevity gains. We conclude that the recent U.S. mortality crisis should be understood not only as a story about particular causes of death, but also as a weakening of institutional and social translation.