Summary
Cutler, Deaton, and Lleras-Muney argue that knowledge and technology — not income — are the ultimate determinants of mortality improvement, both historically and across countries today. They trace three historical phases of mortality decline (improved nutrition, macro public health infrastructure, post-1930s medicine), show that the cross-country income–life expectancy relationship (the Preston Curve) has shifted upward over time due to technology diffusion rather than income growth, and argue that within rich countries the education gradient in mortality reflects the "fundamental causes" mechanism: educated people adopt new health technologies faster, so the gradient persists through successive waves of health innovation.
Key Claims
- The Preston Curve (income vs. life expectancy across countries) has shifted upward dramatically across the 20th century. Income explains only ~10–15% of life-expectancy (LE) changes over 10–40 year periods in cross-country time-series regressions; the remainder reflects autonomous knowledge and technology diffusion independent of growth.
- Three phases of mortality decline: (1) improved nutrition and agricultural productivity (mid-18th to mid-19th century); (2) macro public health — germ theory, water filtration, sanitation (late 19th – early 20th century); Cutler and Miller (2005) estimate water purification alone accounts for ~50% of US urban mortality reduction in the first third of the 20th century; (3) post-1930s medicine — antibiotics, then cardiovascular disease (CVD) treatments; CVD decline accounts for ~70% of the 7-year US life expectancy gain from 1960–2000, with about two-thirds attributable to medical advances and one-third to smoking reduction.
- Within poor countries, cheap life-saving interventions go largely unused. The binding constraints are institutional capacity and political will, not income. India and China both show that rapid gross domestic product (GDP) growth does not automatically reduce infant mortality without deliberate public health investment — growth is neither sufficient nor necessary for health progress.
- Within rich countries, the education-mortality gradient is best explained by the "fundamental causes" mechanism (Link and Phelan 1995): education maintains a health advantage regardless of which specific technology is currently reducing mortality, because educated people adopt new health technologies faster. The gradient therefore persists and adapts rather than being eroded by any single health innovation.
- Psychosocial stress and allostatic load (the Marmot/Whitehall mechanism) provide an alternative pathway for the within-country socioeconomic status (SES) gradient: chronic subordination stress elevates cortisol and inflammatory markers, raising CVD and all-cause mortality risk independently of behaviors or medical access.
- The fetal origins hypothesis (Barker 1992): in-utero nutritional deprivation programs adult cardiovascular disease and metabolic disorder, making early-life conditions a distinct determinant of adult mortality independent of adult income.
- Global mortality structure 2002 (WHO): 30% of deaths in low-income countries are children under 5; in high-income countries, only 0.9% — the contrast compresses the entire historical transition into a single cross-section.
- Income is partly endogenous to health (sick people earn less), making ordinary least squares (OLS) income-mortality regressions upward-biased. True causal income effects are likely smaller than cross-sectional associations suggest.
Concepts Introduced or Extended
Entities Mentioned
Quotes
"The proximate determinants of health are the factors that directly cause mortality — what kills you. The ultimate determinants control the proximate determinants."
"Knowledge, in the form of germ theory, or of the relationship between smoking and cancer, is the key input. Once knowledge exists, it can spread to all countries regardless of income."
"The answer lies in the knowledge and the willingness of governments to implement that knowledge. These are political and social issues, not purely economic ones."
My Take
A superb Journal of Economic Perspectives (JEP)-style synthesis — accessible, provocative, and well-organized. The "knowledge not income" thesis is convincing in broad strokes: the upward shift of the Preston Curve is a powerful stylized fact, and the Cutler-Miller water purification evidence is about as clean as historical causal identification gets. Two weaknesses: (1) "knowledge" is itself endogenous to income — generating, codifying, and disseminating public health knowledge requires investment, which requires wealth or aid. The paper treats knowledge as manna from heaven once discovered, rather than examining why the same knowledge is adopted faster in some poor countries than others. (2) The within-country SES gradient section juxtaposes the fundamental causes and psychosocial stress mechanisms without adjudicating between them, even though they have different policy implications — if the gradient is about technology adoption, innovation-focused policy helps; if it's about chronic stress, redistribution and labor conditions matter more. On developing countries, the paper is descriptively informative but offers no mechanism model for why institutional capacity varies.