Use the Table of Summaries to link to the summaries either by author or by risk factor.
Abstract | Bibliography | Introduction | Table of
Summaries
Income Inequality And Mortality In 14 Developed Countries
Lobmayer, Peter and Richard G. Wilkinson, 2000. Sociology of Health & Illness,
22:401–414. †
The authors have undertaken previous research which showed that more egalitarian
societies tend to have lower mortality rates. However, there have been suggestions that this relation no
longer exists and this research seeks to discover reasons why it is no longer apparent.
There have been changes in the extent of income inequality and in the age distribution of
the relatively poor. The authors propose that this shift in relative poverty from elderly people (with high
death rates) to young families (with low death rates) may obscure the association between income
distribution and mortality. They also wanted to compare the importance of absolute income with income
inequality as predictors of mortality. Data from wave 3 of the Luxembourg Income Study (LIS) were used.
These data relate to the period 1989–92 and provides age– and sex–specific mortality data
from 15 OECD countries and corresponding data on income inequality, median income and absolute and relative
poverty. (The authors only used data from 14 countries; they omitted Luxembourg due to its small size.)
Six categories of age were used, infants (0–1), pre–school (1–4),
children (5–14), early adulthood (15–29), middle–age (30–64) and older people
(65–84). The authors calculated two summary measures of mortality, an age–adjusted death rate
(over all ages) and an age–adjusted rate of Potential Years of Life Lost, (PYLL), for those aged below
65. PYLL is a measure of premature mortality which gives more weight to deaths at younger ages. Income was
measured as 'personal disposable household income' and adjusted for the number of people in each household
using equivalence scales. Both the OECD equivalence scale and a subjective equivalence scale were used as
they make very different assumptions. Income distribution was measured at the 50th and 10th percentiles, i.e.
at the median and at the top of the lowest decile.
An age–specific measure of relative poverty was determined for 12 countries. Data
were inadequate for Sweden and Australia. Relative poverty was measured as the proportion of people in an age
group who were in the bottom decile of their country's income distribution. Absolute poverty was set at the
same level of real income for all countries, using purchasing power parities taken from OECD National
Accounts.
It was found that the two equivalence scales (OECD and subjective) were strongly
correlated and the results based on each showed little difference. As a result only the results based on the
OECD equivalence scale are presented.
Both higher median income and greater inequality were associated with higher premature
mortality. This suggests that wider income distribution is related to higher premature mortality and higher
age–specific mortality rates below age 65. However, these results did not extend for ages over 65. The
relations with age–adjusted mortality, over all ages, were either non–existent or weakly inverse.
The authors suggest that the fact that the measure of premature death excludes the older ages and puts more
weight on deaths at younger ages might account for this. As most deaths occur in old age, the negative
associations between income distribution and mortality in the 65–84 age group dominate the age–
adjusted mortality rate.
There was a close positive relation between income inequality and median income. However,
the authors comment that this relation should not be regarded as an indication that higher median income is
bad for health. They state that this may instead reflect the conflicting effect of income inequality on
mortality.
Age–specific relative poverty was positively related to mortality rates in the
oldest and in the two youngest age groups, but was weakly negative in the other age groups. The authors
suggest that an association with societal inequality and/or reduced relative poverty among people over
retirement age may explain this anomaly. No consistent relation was found between absolute income levels and
mortality.
The USA, having higher and more unequal incomes, was an outlier with higher absolute
income and higher income inequality. This is not necessarily better or worse, it simply highlights that the
USA is very different from the other OECD countries. As a result, the USA lay apart from the other OECD
countries; thus it could make or break relations. The authors confirmed that the data were of a high standard
and therefore the USA could not be treated as an invalid data point. They commented that the US differed in
degree rather than kind such that if the single observation for the US was replaced by data for each of the
50 states, they would form part of a continuum with the more egalitarian states merging with the other
countries.
The authors provide three possible explanations for the apparent disappearance of the
relationship between income inequality and mortality. Firstly, the shift of relative poverty from older
people to young families with children affects the impact of income inequality on morality rates standardized
across all ages. Secondly, the effects of income inequality may be partly confounded by higher median
income. Finally, there is the possibility that there are lagged health effects of changes in income
distribution.
† This study is based on data from 14 OECD countries: Australia, Belgium, Canada,
Denmark, Finland, France, (former West) Germany, Italy, the Netherlands, Norway, Spain, Sweden, the United
Kingdom and the United States.