Has rising income stopped translating into longer life expectancy?
Core argument: U.S. Preston curves relating life expectancy to log real per capita income shifted upward across states from 1980 to 2010 and across counties from 2000 to 2010, but stopped shifting upward between 2010 and 2019 even as incomes kept rising.
Figure 1 shows Preston curves relating life expectancy to log real per capita income across U.S. states and counties. Among states, the curves follow the classic pattern from 1980 through 2010: states became richer and longer-lived [as did counties from 2000 to 2010]. That pattern changed after 2010. Between 2010 and 2019, states and counties continued to become richer, but the curves did not shift upward: resource accumulation continued, but commensurate longevity gains did not. The post-2010 period involved both decoupling and divergence: rising income no longer translated into broad longevity improvement, and places became more unequal in their capacity to convert aggregate resources into longer lives. Figure 4 asks a simple counterfactual question: if earlier covariate–life expectancy relationships had persisted, how much life expectancy would later years have achieved, given the actual changes in these covariates? Together, the decomposition results sharpen the central interpretation. The 2010s mortality crisis did not occur because income, education, wealth, or insurance coverage simply moved in the wrong direction. In many respects, aggregate socioeconomic conditions improved. Instead, what changed was the relationship between those conditions and longevity. The United States became less successful at converting collective resources into population-level longevity gains.