Simulating how income inequality within a population widens the gap in life expectancy between rich and poor — after Chetty et al.
Every analysis of health inequality begins with the same simple act: sorting people by income. Raj Chetty and colleagues did this at unprecedented scale in their 2016 JAMA study "The Association Between Income and Life Expectancy in the United States, 2001–2014," linking 1.4 billion person-year observations from de-identified tax records to Social Security Administration death records. The resulting income ladder — from the poorest to the richest Americans — became the backbone for measuring how much lifespan tracks with wealth.
Comparing raw dollar incomes across decades and regions is misleading — a $40,000 income means something different in rural Mississippi than in San Francisco. Chetty et al. instead ranked every individual by their percentile position within the national household income distribution for a given year. This percentile-rank approach has three advantages:
• It is robust to inflation and regional cost-of-living differences • It captures relative deprivation — a well-documented driver of stress physiology independent of absolute income • It allows direct comparison to the shape of the overall income distribution (captured by the Gini coefficient)
The Gini coefficient summarizes how unequal that ladder is: 0 would mean everyone has identical income, 1 would mean one person has all of it. The United States sits around 0.41–0.49 depending on the year and measure used — among the highest of any wealthy democracy, and rising for the past four decades.
In this simulation, the population is laid out along the horizontal axis by income percentile (0 = poorest, 100 = richest). Every dot represents a slice of the population. As you increase the Gini slider, the underlying income distribution — and everything that flows from it — becomes more stratified.
Between 1980 and 2016, the income share held by the top 1% of US earners roughly doubled, while median wages for the bottom 50% grew by less than 1% per year after inflation. The income ladder did not just get longer — its rungs spread further apart.
Income does not cause longer life directly. It buys — or fails to buy — a set of intermediate resources that accumulate biological wear over decades: consistent healthcare access, nutrient-dense food, safe housing, and lower chronic stress exposure. These resources are distributed almost as unevenly as income itself, and the gaps compound across a lifetime rather than resetting each year.
Chetty et al. and the broader health-inequality literature identify three overlapping channels linking income to biological aging:
1. Healthcare access — insurance coverage, primary-care continuity, and time-to-diagnosis for treatable conditions (hypertension, diabetes, early-stage cancer) all improve steeply with income. Low-income adults are more likely to delay care until a condition becomes an emergency.
2. Environmental and material exposure — housing quality, air pollution, occupational hazard exposure (manual labor, night shifts), neighborhood safety, and food-desert access to fresh produce all correlate strongly with income percentile. These are "upstream" exposures that operate for decades before any diagnosis appears.
3. Health behaviors — smoking, obesity, and physical inactivity rates all show a strong income gradient, though Chetty's team found behaviors explain only part of the mortality gap; local-area health infrastructure and stress exposure matter independently.
Critically, low income does not act like a light switch — its effects are cumulative. A child raised in a low-income household inherits worse nutrition, more air pollution, and more chronic stress before ever earning an income themselves, and these early exposures shape adult disease risk (the "biological embedding" of early-life disadvantage).
When unequal healthcare access, environmental exposure, and chronic stress compound over 40+ years of adult life, they surface as a large, measurable gap in life expectancy. Chetty et al. found that the richest 1% of American men live on average 14.6 years longer than the poorest 1%; among women the gap is 10.1 years — differences comparable to the mortality impact of smoking versus never smoking.
The life-expectancy gap is not a single event — it is the integral of small mortality-risk differences accumulated year over year across the population. In this simulation, the population "band" starts near-flat at the national average life expectancy (78.5 years) and stretches apart over time as unequal resource access compounds: the top-percentile curve bends upward while the bottom-percentile curve bends downward, forming a widening envelope around the mean.
This mirrors the real epidemiological pattern: mortality-risk gradients by income are modest at any single age, but because they apply every year of adult life, they sum into a double-digit gap in expected lifespan by the time a cohort reaches its 70s and 80s. The slope of the divergence — how steep the gap grows per percentile of income — is set by the Gini coefficient: more unequal income distributions produce steeper, faster-diverging life-expectancy curves.
Chetty's team found that low-income life expectancy varies enormously by where people live (up to 4.5 years), while high-income life expectancy is almost the same everywhere — suggesting local policy and environment matter most for the poor, not the rich.
One of Chetty et al.'s most striking findings is that the income–longevity gradient is not uniform across the country. Low-income residents of high-cost, high-investment metro areas (e.g., San Francisco, New York) live meaningfully longer than equally low-income residents of struggling post-industrial or rural regions — even though their incomes are identical. Local factors such as public health infrastructure, smoking norms, and social cohesion appear to matter enormously for the poor, and far less for the rich.
Chetty et al. correlated local-area life expectancy for low-income residents against dozens of community characteristics and found the strongest associations with: local smoking prevalence, obesity rates, and measures of social capital and civic engagement — not with local medical spending or insurance coverage rates per se. Areas with more college graduates, higher local government expenditure, and greater income mixing (less residential segregation by class) show systematically longer lifespans for their low-income residents.
High-income residents, by contrast, show almost no variation in life expectancy across geography — they can access quality healthcare, safe housing, and healthy food virtually anywhere, insulating them from local conditions. This asymmetry implies that place-based interventions — improving a community's health environment rather than just individual incomes — could be a powerful lever specifically for reducing the bottom of the distribution, without needing to change the income distribution itself.
If unequal income produces an unequal lifespan, can policy reverse it? Two broad levers exist: reducing income inequality itself (progressive taxation, minimum wage, earned income credits) and investing directly in the health-relevant resources of low-income communities (community clinics, smoking-cessation programs, food access, housing quality) regardless of the income distribution. Simulation and empirical evidence both suggest the second lever narrows the mortality gap faster per dollar spent.
Income redistribution changes the shape of the income ladder itself — flattening the Gini coefficient — but its effect on life expectancy operates on a long, multi-decade timescale, because today's adult mortality reflects exposures accumulated over an entire prior lifetime. Nordic countries with Gini coefficients near 0.27 (versus the US at ~0.41) do show substantially smaller income–longevity gaps, consistent with this pathway, but achieving that shift requires sustained multi-generational policy commitment.
Targeted health investment — expanding Medicaid-style coverage, community health workers, smoking-cessation and nutrition programs concentrated in low-income areas — acts on the proximate causes of the gap directly and can show measurable mortality effects within a decade, because it intervenes on the actual mechanism (healthcare access, stress, behavior) rather than waiting for income effects to filter through. In this simulation, increasing "Targeted Health Investment" compresses the bottom of the life-expectancy band toward the population average far more efficiently, per unit of the slider, than an equivalent reduction in the Gini coefficient alone.
Chetty et al. modeled several policy levers and found that a 1-percentage-point increase in local health-related spending targeted at the bottom income quartile closed more of the life-expectancy gap than a proportionally larger, broad-based tax-and-transfer shift — because the intervention lands directly on the resource bottleneck rather than passing through it.