|
|
This is only a preview of the paper
Click here to register and get the full text.
Existing members click here to login
|
|
|
... INTRODUCTION
The nature of the relationship between income and health has interested researchers for decades.
Somewhat surprisingly, this has not been the case for the relationship between wealth and health. We find
this surprising because wealth may be a better measure of a person’s economic resources than current
earnings. The analysis of the link between income and health suggests a positive but declining marginal
relationship between the two (i. ... , the gradient between income and health is steepest among lowerincome
individuals). Does the same relationship exist between health and wealth?1 And does the
suggested relationship between income and health, as first illustrated by Samuel Preston (1975), hold up
once wealth has been taken into account? The few studies of health and wealth have in general been
limited to older populations, particularly those sampled in the longitudinal Health and Retirement Study
(HRS) and Assets and Health Dynamics among the Oldest Old (AHEAD) surveys. The findings from
these studies have suggested that among those nearing or past retirement age, greater wealth and better
health are dynamically related, and that this link is stronger than the link between income and health. In
this paper we explore the relationship between health and wealth and health and income for prime
working-age individuals, those between the ages of 25 and 54, in the United States.
The influence of wealth on health may occur through a variety of channels. Savings may provide
crucial resources when an individual is struck by an unexpected health shock. For individuals with little
disposable income, lack of wealth may limit the ability to purchase potentially effective care, lead to
delays in seeking care, or place them in other situations with additional health risks. Identifying such a
1Limited previous research has suggested that health and certain dimensions of wealth are related among
U. ... Robert and House (1996) find such a link between various health outcomes and having
$10,000 or more in assets; Drentea and Lavrakas (2000) find a negative relationship between credit card debt and
health. ... )
2
health-wealth link would provide individuals a strong incentive to save. This would be particularly
relevant for the working-aged population who are still accumulating wealth. The study of the healthwealth
relationship is also relevant to many questions asked by researchers of income and health. Income
inequality may cause stress and thus poor health, and wealth inequality may do so as well. If health
affects employment outcomes, then it is likely to affect an individual’s long-term savings or wealth.
Finally, understanding these links may give us some prediction of whether to expect increasing health
differences as income and/or wealth inequality increases.
Although study of the causal mechanisms of health-wealth dynamics is necessary to answer these
questions, characterizing the cross-sectional relationship between health, income, and wealth can provide
information that supports or rejects certain hypotheses, or that raises other interesting research questions.
Victor Fuchs (1982), for example, has suggested that “third variables” such as time preference may
explain the links between socioeconomic status and health. Time preferences may influence the
attainment of both health and wealth, resulting in the spurious positive correlation between the two. ...
In the next section, we summarize the current literature on the relationship between health,
income, and wealth. ... The results section contains a detailed
description of the cross-sectional relationship between health and wealth among individuals aged 25 to
54. This includes an analysis of health-wealth differences by age and by race group, and a
characterization of the health gradient. ... THE EVIDENCE TO DATE
The literature relating health to socioeconomic status (SES) has a long history (see recent reviews
in Adler and Newman, 2002; Deaton, 2001; Goldman, 2001; Robert and House, 2000a, 2000b). ...
In each of these cases, a strong positive relationship between health and SES has been established. More
recent studies have explored the relationship between health, income, and income inequality (see recent
reviews by Deaton, 2001; Mullahy, Robert, and Wolfe, 2001; Robert and House, 2000b; and Wagstaff
and van Doorslaer, 2000). There is a consensus among these authors that given the research to date, there
is (1) a well-documented concave relationship between individual income and health, but (2) little
evidence that income inequality affects health.
The literature relating health to wealth is more limited. This is partly due to the difficulty in
collecting wealth data and the limited availability of surveys that collect both wealth and health
information in the United States. There is some evidence, however, that wealth and health are positively
correlated, even net of income. Using the Americans’ Changing Lives Study, a cross-sectional national
sample of adults in 1986, Robert and House (1996) found that assets are associated with self-rated health,
functional health, and a number of chronic conditions, over and above income, education, home
ownership, and age, sex, and race. ...
Additional evidence of a health-wealth link net of other SES measures comes from the Whitehall
studies of British civil servants. ... Additionally, Drentea and
Lavrakas (2000) found that credit card debt was significantly related to having both physical impairments
and poor self-reported health.
4
A small number of recent studies have addressed the dynamic nature of the health-wealth
relationship focusing on the HRS and AHEAD samples of older Americans. The mechanisms through
which wealth may be related to health have been described by Goldman (2001).2 Explanations for the
health-wealth link can be categorized into three general types of mechanisms: ways in which health
affects wealth; ways in which wealth influences health; and links between health and wealth that are due
to other common correlates such as age or differences in individual time preferences (discount rates).
Smith (1999) used data from the HRS, limited to individuals aged 51 to 61, to consider the effects
of health shocks on individual wealth. Smith found a strong effect of new illness on both income and
wealth. ... Their study of the causal relationship between wealth and health
rejected the hypothesis that there are causal effects of SES on health when health is measured by acute
illnesses. However, they could not reject the hypothesis that SES causes health conditions with gradual
onset, such as mental illnesses, or that health shocks affect wealth.
Smith and Kington (1997) also analyzed the AHEAD sample and found (1) a strong nonlinearity
in the relationship between health and both wealth and income—they solved the nonlinearity problem by
estimating separate spline linear effects for income and wealth terciles; (2) both income and wealth were
related to health in this elderly population; (3) different types of income were differentially related to
health; (4) race and ethnic differences in health were explained by income and wealth for functional
limitations, but not other health outcomes.
The wealth accumulation research has indirectly addressed health effects over a broader range of
age groups. For example, Altonji and Doraszelski (2002), Shea, Miles, and Hayward (1996), and Smith
2Other heuristic discussions regarding health effects of wealth and wealth effects of health can be found in
Kotlikoff (1989), Smith (1999), and Smith and Kington (1997).
5
(1995), among others, have consistently found a significantly positive relationship between wealth and
health among whites. ... 3
In summary, most of the evidence regarding the relationship between wealth and health suggests
a positive nonlinear association for whites. Among the working-age population, there is some evidence
that assets and debt are associated with health over and above income and other measures of SES. ... The research described above also suggests that the
relationship between wealth and health may be bidirectional. Our review finds that little is known about
the characteristics of the health-wealth relationship among younger populations, and whether it resembles
the gradient relationship between health and income, particularly once income is held constant. ... adult population using rich
wealth data, distinguishing between income and wealth as they relate to health. ... Compared to other
national surveys, the SCF stands out for its measurement of wealth—wealthy individuals are oversampled
so that accurate estimates can be made at the highest income and wealth levels. The SCF is considered to
be the gold standard in terms of capturing the distribution of wealth in the United States. ... Along
with detailed information regarding household assets, the SCF collects demographic, health, and income
data from all survey participants. In this study we use data aggregated from the 1989, 1992, 1995, and
3For additional examples of studies including health as an independent variable in wealth equations, see
Diamond and Hausman (1984), and Avery and Rendall (1997). ...
4See Smith (1995) and Wolff (1999) for a comparison of the wealth measure in the SCF to those in other
surveys. ... Measurement of Health Status
The SCF collects basic information regarding health. Each survey year, the following question is
asked of respondents: “Would you say your health is excellent, good, fair, or poor? ... Although the four-category
scale has been shown to be culturally sensitive, a dichotomous measure reflecting a response of “poor” or
“fair” health has been shown to be both reliable and highly correlated with other measures of morbidity
and mortality (Baker, Stabile, and Deri, 2001). We use this dichotomous self-report of poor/fair health in
our study. ... 5 percent of our weighted sample of household heads reported poor or fair
health.
A change in the survey questionnaire between 1992 and 1995 had a possible impact on the fourcategory
measure of self-reported health. ... This new smoking question immediately preceded the selfreported
general health question, possibly resulting in the notable decline in reports of excellent health in
1995. It appears that this change had a much smaller effect on the collapsed dichotomous measure of
poor/fair health. Nevertheless, year dummies included in the models of poor/fair health should be
interpreted as including any effect of the survey change.
5Among couples, this implies that only the health of the household head is considered in analyses. Health
and demographic information are also collected for the spouse or partner in married or cohabiting households. ... Measures of Wealth and Income
We use net worth of the family to measure wealth. ... It does not include defined benefit pension wealth,
defined contribution pension wealth held outside 401(k)s, social security wealth anticipated at retirement,
consumer durables, and future earnings (see Scholz and Levine, 2004, for more detail on this measure). ... Like our wealth measure,
total family income is adjusted for family size and the CPI-U.
The use of a family-size-adjusted version of income and wealth requires some further discussion
and justification. ... The use of an
equivalence scale reflects the assumption that when income is used to purchase consumption goods,
including health prevention and maintenance, it must be shared between family members. Similarly, we
assume that if families save for precautionary health purposes, they do so for all members in the family.
In the case of a large health shock, although it can be argued that the entire unconsumed portion of family
income and wealth may be made available to the individual, other household consumption needs will still
constrain these resources. ...
Distributional characteristics of wealth and income, both unadjusted and adjusted for the
household equivalence scale, are shown in Table 1.6 On average, total family income and wealth are
approximately twice the size of income and wealth adjusted for family size. ... Demographic and Other Explanatory Variables
Existing evidence suggests that the relationship between wealth and health differs by age and
race. ...
9
Table 1
Weighted Wealth and Income Summary Statistics (Standard Errors)
among Household Heads Aged 25–54
Total Family
Income or Wealtha
Family Income and
Wealth Adjusted for
Family Sizea
Income
Average income $58,440 (1,370) $31,290 (766)
10th percentile 10,516 (221) 5,619 (180)
25th percentile 22,072 (377) 11,694 (240)
50th percentile (median) 40,758 (591) 21,808 (329)
75th percentile 66,083 (1,201) 35,395 (557)
90th percentile 100,723 (3,026) 56,337 (1,426)
Wealth
Average net worth $189,921 (11,353) $99,604 (5,817)
10th percentile 0 (0.
Approximate Word count = 9499
Approximate Pages = 38
(250 words per page double spaced)
|
|
|
|
|
|