Incomes and Mortality Patterns

01/01/2002
Summary of working paper 8534
Featured in print Digest

There is no evidence that changes in income inequality affect changes in mortality. Adult and elderly mortality rates declined most rapidly during the period when income inequality increased.

A new study of mortality rates in Britain and the United States finds no simple relationship between income growth, patterns of income inequality, and the substantial declines in mortality rates in the two countries since 1950. In Mortality, Income, and Income Inequality Over Time in Britain and the United States (NBER Working Paper No. 8534), NBER Research Associates Angus Deaton and Christina Paxson cast doubt on the idea that income trends explain mortality trends in the post-war period, and that rising income inequality is associated with higher mortality. Rather, the researchers highlight the importance of advances in medical technology in explaining death rates as well as the important differences in the health care systems in the two countries.

Deaton and Paxson exploit the similarities and differences in patterns of income in the United States and Britain. The two countries have experienced different patterns in income growth: the growth rate of family income in the United States was high from1950 to the early 1970s, low during the 1970s and 1980s, and high again in the 1990s. In contrast, incomes in Britain grew fairly steadily during the second half of the twentieth century, with little slow down in the 1970s and 1980s.

Although these income growth patterns differ, changes in income inequality are similar in the two countries. Inequality changed relatively little until the early 1970s; it then started to increase in the United States and, in the late-1970s, in the United Kingdom. The increase in income inequality accelerated in both countries in the mid-1980s, and leveled off in the 1990s.

Despite sharp differences in patterns of income growth, the two countries exhibited similar patterns in mortality decline. Furthermore, the sharpest declines in mortality were coincident with the productivity slowdown in the United States. Therefore it is not plausible that mortality declines are driven by changes in income levels. Furthermore, there is no evidence that changes in income inequality affect changes in mortality. Adult and elderly mortality rates declined most rapidly during the period when income inequality increased.

The patterns of declining mortality in both countries vary across age and gender groups. Men always have higher mortality rates than women in the same age group. Younger groups experienced a more rapid mortality decline before 1970, and much less rapid mortality declines and, in some cases, notably for young men, mortality increases, in the later years. Deaths from AIDS account for almost all of this increase in mortality for young men. For those over age 40, there is a rapid decline in mortality throughout the period with the trend accelerating after 1970. British mortality is lower for those younger than age 65, but the schedules cross at age 65, after which mortality rates are lower for elderly U.S. citizens.

If changes in incomes and income inequality do not explain mortality patterns, then what does? Two likely candidates are technological advances in health care and differences in the health care systems in the two countries. The researchers show that declines in mortality in the United States are mirrored by mortality declines in Britain that occur after a lag of about four years. It may be that medical innovations are introduced first and diffuse more quickly in the United States than in Britain. The authors speculate that the centralized health care system in Britain may impede the adoption of expensive new technologies. In the competitive U.S. health care industry, there may be greater pressure to adopt new technologies as soon as they are feasible, regardless of cost.

Differences in the health care systems also may be responsible for differences in the age patterns of mortality in the two countries. The United States has largely private provision of health care until the age of 65, but not all citizens have access to health services until they are eligible for Medicare. Although Britain has universal access at all ages, less is spent on health care than in the United States, and access to health care is rationed, particularly by age. British elderly are frequently denied access to expensive technologies from which they are likely to benefit. These institutional features are consistent with the observed patterns in mortality, in which the U.S. mortality is higher than Britain's for younger age groups, but lower after retirement. Because this pattern was established well before the Medicare program began, and because the differences in mortality rates between the two countries begin to narrow with age prior to Medicare eligibility, rationing of care by age likely has more effect on the mortality of British elders than does Medicare on American elders.

-- Andrew Balls