Abstract
Rodgers' paper on income inequality as a determinant of population health starts, like so much else, from Preston's paper on international patterns of income and health, published in the same journal four years previously.1, 2 Preston showed that life expectancy increased with income across countries, but at a rate that became progressively lower as income increased— there are diminishing health returns to income—and noted that, if similar relationships between income and health held within countries, a country with a more equal distribution of income would have higher life expectancy, other things being equal. Rodgers tested Preston's conjecture using a sample of 56 (unnamed) developed and developing countries, and found, indeed, that the Gini coefficient of income inequality had a significant negative effect in a (non-linear) relationship between average income and life expectancy (and infant mortality). In the life-expectancy regressions, the coefficient on the Gini coefficient varied with the specification, but was around –40. This estimate would imply a difference in life expectancy of 12 years between two countries with identical mean income but with Ginis of 0.6 and 0.3, which are about the extremes that we observe.