Chetty et al. 2016 — The Association Between Income and Life Expectancy in the United States, 2001–2014

Summary

A landmark study using $1.4$ billion de-identified federal tax and Social Security Administration death records (1999–2014) to estimate period life expectancy at age 40 by household income percentile, sex, and commuting zone. Finds a steep, continuous income–life expectancy gradient; widening inequality in longevity over time concentrated at the extremes; enormous local area variation in low-income life expectancy; and health behaviors (not healthcare access or labor market conditions) as the dominant correlate of geographic variation.

Key Claims

• The gap in life expectancy between the top $1\%$ and bottom $1\%$ of the income distribution is $14.6$ years for men and $10.1$ years for women (pooled 2001–2014). The gradient is continuous — there is no income threshold above or below which additional income stops mattering.
• Between 2001 and 2014, life expectancy increased $2.34$ years (men) and $2.91$ years (women) in the top $5\%$ of the income distribution, but only $0.32$ years (men) and $0.04$ years (women) in the bottom $5\%$. Inequality in longevity is widening.
• For individuals in the bottom income quartile, life expectancy differs by approximately $4.5$ years between commuting zones with the highest and lowest longevity. The worst–life expectancy (LE) commuting zones — Gary IN, Detroit MI, Toledo OH, Oklahoma — are the same rust belt and Appalachian regions with highest Disability Insurance (DI) incidence.
• Geographic variation in low-income life expectancy is strongly correlated with health behaviors (smoking $r = -0.69$, obesity $r = -0.47$) but not significantly correlated with healthcare access, income inequality, or local labor market conditions.
• Top $1\%$ men claim Social Security and Medicare for $11.8$ more years than bottom $1\%$ men; top $1\%$ women claim $8.3$ more years. The progressive benefit structure is substantially offset by the life expectancy gradient.
• Indexing Social Security and Medicare eligibility ages to life expectancy gains would be regressive unless conditioned on income and location, because LE gains have been concentrated at the top of the distribution.

Concepts Introduced or Extended

• Income-Mortality Gradient
• Period Mortality
• Deaths of Despair

Entities Mentioned

• Raj Chetty
• David Cutler
• Social Security Administration
• Disability Insurance Program

Quotes

"Men and women in the top 1% of the income distribution can expect to claim Social Security and Medicare for 11.8 and 8.3 more years than men and women in the bottom 1% of the income distribution."

"Men in the bottom 1% of the income distribution at the age of 40 years had an expected age of death of 72.7 years... equivalent to the mean life expectancy for 40-year-old men in Sudan and Pakistan."

My Take

The paper's most underappreciated finding for DI purposes is not the national gradient (which was known) but the local area variation: the same geographic belt — rust belt, Appalachia, Oklahoma — that produces the most DI applications also produces the worst life expectancy for low-income individuals. This geographic coincidence is not incidental. It reflects a common underlying driver: concentrated low-income populations, high rates of smoking and obesity, and the musculoskeletal and cardiovascular conditions that feed both early mortality and DI incidence. The social insurance redistribution finding also reframes DI: because DI beneficiaries are at the very bottom of the income-LE gradient, cash DI benefits reach people in the worst-LE tier, making DI more genuinely redistributive than Social Security retirement benefits alone.