Use the Table of Summaries to link to the summaries either by author or by risk factor.
-
Abstract | Bibliography | Introduction | Table of
Summaries
Income
Knox and Tomlin (1997) found that there was very strong evidence to suggest that
mortality rates were not equal at different income levels. Kallan (1997) concludes that income and employment
status have significant net effects on both all–cause mortality and cause–specific mortality for
all age–sex groups, with the exception of older females. Montgomery and Pappas (1996) found that
poverty had the strongest consistent effect on children's health. Rogers et al (1999) comment that family
income is a primary link between education and mortality. The inverse relation between income and mortality
has been identified, regardless of sex, age, race or marital status, and is apparent even after
adjustments for variations in education, occupation, health status and other demographic characteristics.
Pappas et al (1993) also found that not only did poor people have higher death rates than wealthier
people, death rates fell consistently as levels of income increased. The mortality disparity has increased
between 1960 and 1986. Similar findings are reported in Attanasio and Emmerson (2001), Williams and
Collins (1996) Sorlie et al (1995), Deaton and Paxson (1999), Wolfson et al and the Statistical Bulletin
(1975).
In Judge (1995), Wilkinson states (in the commentary) that income is a determinant and
indicator of a wide range of material factors that not only affect standard of living but also have a crucial
impact on psychosocial factors such as sense of control, security and status. He highlights that the
association between income and mortality is independent of fertility, maternal literacy and education in
developing countries and of average incomes, absolute levels of poverty, smoking, racial differences and the
provision of medical services in developed countries.
It has been argued that rather than income affecting health, health may impact on income.
However, Deaton and Paxson (1999) show that only some of the effect of income was removed when allowing for
reverse causality. Wolfson et al (1990) agree.
Deaton (1999) concluded that income exerts a strong protective effect, with permanent
income being more protective than current income. Wolfson et al (1990) has similar findings. In fact, Deaton
and Paxson (1999) noted that the effects of income are dynamic, and although permanent income is negatively
correlated with mortality, transitory income was positively correlated to mortality and that it positively
followed the business cycle. They suggest that long–run income was a better predictor of health than
current income. Wolfson et al also found that permanent rather than transitory earnings is the key variable.
Thus earnings have long term effects on mortality. Indeed, they found that higher earnings in late middle age
were associated with significantly lower mortality two decades later.
Rogers (1995) comments that there are substantial changes in income for those over 65. He
argues that the lower mortality rates of married couples may be attributable to the economic advantage of
double income. Wolfson et al (1990) found that while there are clear mortality gradients by income within
virtually all age ranges, the relationships are not as strong for those over age 65. However, Knox and
Tomlin (1997) found that for those aged 60–79 mortality was still related to income, with a trend
towards lower mortality rates as income increases. They highlight that the disparities in death rates
between high and low income earners were not as pronounced for the older age groups. The disparities
diminished markedly after age 74. Deaton and Paxson (1999) found income mattered most in middle age
(40–54) and also that the effect diminished with age. Wolfson et al (1990), Williams and Collin
(1995), Sorlie et al (1995), and Rogers et al (1999) have similar findings. They highlighted a striking
positive correlation between income and mortality among young men, which was not observed for young women.
Both Knox and Tomlin (1997) and Vaillant and Mukamal (2001) find that the mortality and
the health status of persons over 80 are not significantly related to pre–retirement income.
Deaton and Lubotsky (2001) state that blacks have higher mortality rates than whites and,
on average have lower incomes. Williams and Collins (1995) also comment that gains in economic status have
stagnated for blacks. In contrast, Montgomery and Pappas (1996) found that the effect of poverty on
children's health did not differ by race.
Rogers et al (1999) comment that those with lower incomes were more likely to be smokers
and less likely to quit. However, in their study, income did not alter the effects of smoking significantly.
Lantz et al (1998) found that the distribution of four behavioral risk factors (cigarette smoking, alcohol
drinking, sedentary lifestyle and relative body weight) significantly varied by annual household income.
Those with the lowest income were significantly more likely to be current smokers, overweight and in the
lowest quintile for physical activity. However, income was predictive of mortality, rather than strongly
related to health behaviors.
Bucher and Ragland (1995) found that income was inversely related to blood pressure,
cholesterol and smoking, but none were significant. The correlation of income with age and height was significant. Income was significantly related to mortality from all causes and lung cancer.
Pappas et al (1993) noted that over time absolute death rates declined for people of all
income levels, but the reduction was greater for those with higher incomes. This highlighted the increasing
disparity in mortality rates for those of different income levels.
Williams and Collins (1996) mention that it has been suggested that mortality rates are
not strongly related to a country's overall economic status, but to the level of income inequality within the country.
Several authors also investigate the connection between mortality and income
distribution.
Wilkinson (1997) argues that the relation between individual income and mortality is
primarily an effect of relative income. He states that absolute income levels are no longer important in the
developed world, rather there is a relation with income distribution, whereby health is worse when there is
greater inequality across the social gradient. He does acknowledge that within a small area of analysis,
income is related to mortality and income distribution is not, due to the lack of heterogeneity in a small
area. Kennedy et al (1996) also suggest a significant positive relation between income distribution and life
expectancy and infant mortality. These results were robust to adjustments for poverty, median household
income and household size. Kennedy et al found that the relation of income inequality to total mortality was
higher for the black population but comment that the association between income inequality and mortality is
not completely explained and that income distribution may be a proxy for other social indicators.
This last comment is supported by Judge (1995). In this paper, Judge refutes Wilkinson's
initial claims that income inequality is the key determinant of variations in average life expectancy among
developed countries. He views variations in income distribution to be the product of many influences such as
dietary influences, cultural factors and national perceptions of self-–esteem, that interact over long
periods of times.
In a later paper, Judge et al (1998) reach the same conclusion. However, the authors
state that the existence of a link between income inequality and average levels of population health is a
plausible proposition and comment that they cannot confidently state that income inequality is not associated
with national health. They considered whether changes over time in income inequality and population health
are associated with each other and although the coefficients for life expectancy were in the expected
direction, none were significant. They highlight that any link between income inequality and health is
exposed to the risk of confounding variables.
Deaton (1999) also challenges Wilkinson's hypothesis and analyses. Higher income was
shown to lead to lower mortality and higher income inequality was shown to lead to higher mortality. However,
the effects were not robust when the mortality was age standardized. Nor did Deaton (1999) find a clear and
consistent correlation between inequality and life expectancy when tracking on a consistent basis over a
period of years. He concluded that there was no evidence that higher income inequality raises the risk of
dying. He suggests there is a risk of spurious correlations being made with income inequality being a proxy
for a wide range of social ills.
In contradiction to the Wilkinson hypothesis, Deaton and Paxson (1999) actually found a
protective effect whereby higher inequality is associated with lower mortality.
A further paper by Lobmayer and Wilkinson (2000) found that both higher median income and
greater inequality were associated with higher premature mortality. This suggests that wider income
distribution is related to higher premature morality. The results did not extend for ages over 65 and they
suggest that this is due to a shift in relative poverty from older people to young families with children.
They conclude that there may be a large impact of income inequality on younger lives and that combining all
ages in life expectancy masks this.
The paper notes that the USA, having higher and more unequal incomes, was an outlier
highlighting that the USA is very different from the other OECD countries. A similar result was found in
Judge et al (1998).
In contrast to earlier papers, Deaton and Lubotsky (2001) showed that income inequality
is correlated to mortality for both blacks and whites separately. The fraction of the population that is
black was found to be positively correlated with average white incomes, and negatively correlated with
average black incomes. Once the fraction black was controlled for, there was no relationship between income
inequality and mortality for the whole population or for whites. This implies that the component of income
inequality that matters for mortality is income inequality between races, not income inequality within them.
They concluded that conditional on racial composition, income inequality does not raise the risk of
mortality. They highlighted that the effects of inequality are inconsistent and more often estimated to be
protective than hazardous.
Attanasio and Emmerson (2001) could not determine whether the relevant concept is
elative or absolute wealth.