What the evidence says
Across the sources gathered here, a higher-income population consistently shows lower mortality than a lower-income population, and the gradient holds whether the comparison is by income quintile within a national population or between an upper-income pension-covered subgroup and the national baseline. In a cross-sectional mortality model, the fitted mortality curves are strictly ordered by income quintile across ages 50–120, with the lowest-income quintile showing higher baseline mortality than the highest, and the quintile curves converge (move closer together) at advanced ages. In Brazil, where pension-fund membership is concentrated in the upper deciles of the national income distribution (under 8% of the population holds a pension plan), the pension-covered population and the national population show a mortality gap that widened over the observed 2013–2018 window, with income differences and healthcare access cited as the source of the mortality-rate disparity between occupational subpopulations and the national population in high-inequality countries. Income volatility — a distinct construct from income level — showed no measured effect on all-cause mortality in a population-based cohort (p for trend = 0.847).
A separate strand of evidence, from a Chinese defined-contribution pension system, shows the income-mortality gradient producing a downstream fiscal effect rather than a mortality effect directly: because the system converts an account balance to an annuity using one official divisor regardless of income, but higher-income retirees live longer, the actuarially fair divisor is longer for higher-income quintiles than for lower ones. At retirement age 60 (reference year 2020), the fair counting month ranges from 157.0 months (lowest income quintile) to 161.1 months (highest), against an official value of 139 months, implying subsidy rates of 13.0% to 15.9% across quintiles — a 2.9-percentage-point gap. At age 63, the fair counting month ranges from 153.3 to 157.8 months against an official value of 117, implying subsidy rates of 31.0% to 34.8% — a 3.8-percentage-point gap. Projected to 2040, the income gradient in subsidy rates remains ordered and tends to widen as mortality continues to improve, reinforcing what the source calls a reverse transfer from lower-income to higher-income retirees.
For maternal mortality in China (panel data, 2009–2021), total wealth (the average of urban and rural resident income) decreases the maternal mortality rate: β = −0.00050 (95% CI −0.00068 to −0.00032), with an adjusted R² of 92.3% for the full-sample model. Household income was the only socioeconomic factor that emerged as a consistent driver of maternal health outcomes across all models tested. The proposed mechanism is that income increases household capacity to afford antenatal visits, skilled delivery, and timely emergency treatment, reducing avoidable delays under a "three delays" framework.
Income also decreases wealth-related inequality in maternal mortality in the full China sample, measured by two indices: SII (β = −5.52e-08, p < 0.001) and logRII (β = −0.00110, p < 0.001). Provinces with higher average resident income show a reduced concentration of maternal deaths among poorer regions, and the source frames economic development as increasing income levels alongside more equitable maternal health outcomes.
Coverage is uneven. There is no claim here isolating an income-mortality gradient specifically within a female cohort for all-cause mortality — the female-specific evidence available is limited to maternal mortality. There is also no pension-population evidence outside Brazil (mortality-gap framing) and China (annuity-subsidy framing); no other country's pension population appears in these claims.