INTRODUCTION
The financially better-off also tend to be in better health. This holds between and within countries, both developed and developing, and it has been evident for a considerable period of time (Hibbs, 1915; Van Doorslaer et al., 1997; Woodbury, 1924).
There is an income gradient in mortality, as well as in a variety of measures of morbidity and disability. The income-health relationship is not confined to a health gap between the poor and the rest, however. Health continues to rise with income among the nonpoor.The strength, ubiquity, and persistence of the positive relationship between income and health make it of considerable interest for those studying distributions of income, health, and well-being. Understanding the mechanisms that generate the income-health nexus can help account for inequality, as well as identify inequity, in each of those distributions. This chapter examines the strength and nature of the relationships between income and health with the purpose of establishing the extent to which the distribution of health in a population contributes to economic inequality and is itself a product of that inequality.
The distribution of health is potentially both a cause and a consequence of the distribution of income. Differences in health can generate differences in income, most obviously by restricting earnings capacity. But health inequality may itself reflect economic inequality if health-enhancing goods, such as medical care and nutritious food, are allocated by price. The potentially bidirectional relationship between health and income is relevant both to the positive explanation of the distribution of income and to its normative evaluation. A full understanding of how income differences across individuals are generated requires identification of the extent to which health constrains income. This positive exercise feeds into the normative one of evaluating the distribution of income because the inequity ofincome inequality surely depends on its causes.
The ethicaljudg- ment of income distribution is also contingent on its consequences. If money can buy health, then there may be greater aversion to inequality in the distribution of income than there would be if the rich were merely able to afford smarter clothes and faster cars.The relationship between income and health is not only of interest to those concerned with the distribution ofincome. From the public health perspective, attention is drawn to observed increases in health with income, as opposed to the corresponding decrease in income with ill-health. Public health scientists tend to interpret the income gradient in health as a symptom of inequity in the distribution of health (Commission on the Social Determinants of Health, 2008), while economists are inclined to view the gradient as reflecting the operation of the labor market in which the sick and disabled are constrained in their capacity to generate earnings (Deaton, 2002; Smith, 1999, 2004). Resolution of this debate is obviously crucial to the formation of the appropriate policy response to the gradient. Ifit mainly reflects the impact of ill-health on income, then the proposal to use income redistribution as an instrument of health policy (Commission on the Social Determinants ofHealth, 2008; Navarro, 2001) would be entirely inappropriate (Deaton, 2002).
The inclusion of this chapter in the Handbook is partly motivated by insights into the explanation and evaluation of income distribution that can be gained through the study of the income-health relationship, but it also reflects a trend away from the more narrow focus on differences in income to the more encompassing analysis of inequality in well-being.
Health and income are typically cited as the most important determinants of well-being and are the most common arguments of multidimensional measures of inequality (see Chapters 2 and 3). For given degrees of inequality in the marginal distributions of income and health, most would consider that inequality in well-being is greater when the poor also tend to be in worse health.
Understanding the nature of the relationship between income and health is central to determining the degree of inequality in well-being.17.1.1 Health to Income
There are multiple mechanisms through which health may affect income distribution, with the labor market obviously being an important one. Differences in productivity deriving from variation in physical and mental capacities related to illness and disability are also potentially important determinants of earnings. Differences in the nature of work and infrastructures may mean that physical disability represents a greater constraint on earnings in low-income settings, and mental health problems are relatively more important in developed countries. Discrimination may further widen any disparity in earnings between the disabled and able-bodied. Institutional constraints on wage flexibility may result in unemployment of less healthy individuals who are less productive or face discrimination. On the supply side, ill-health may shift preferences away from work, and this may be reinforced by reduced financial incentives arising from a lower offer wage and entitlement to disability insurance (DI). The latter will cushion the earnings loss arising from disability and so compress the income distribution, but this will be offset if the financial incentives induce withdrawal from employment at a given degree of disability, which may strengthen the earnings—health relationship, if not the income-health relationship, in high-income countries relative to low-income countries. Beyond its effect on the distribution of personal income, health may impact the distribution of household income through the formation and maintenance of marriage partnerships and spousal earnings given needs for informal care.
The impact of health on income may operate with a very long lag. Poor health in childhood may disrupt schooling. Exposure to health risks in utero and illness in infanthood may impair cognitive functioning and reduce the efficiency of education in producing knowledge and skills.
Childhood health problems may be persistent, such that less healthy young adults enter the labor market with less human capital and lower prospects for lifetime earnings. Early-life health conditions may impact income not only through human-capital acquisition but also by triggering health problems in adulthood (Barker, 1995) that interfere with work. If exposure to health risks in early life is related to economic circumstances, then childhood health could be partly responsible for the transmission of these circumstances across generations (Currie, 2009). According to this proposition, poorer mothers with less education deliver less healthy babies and raise sicker children who acquire less human capital and suffer persistent health problems, both of which constrain earnings and increase the likelihood of parenting a child with health problems. If this theory is empirically significant, then it would place health policy at the very heart of social policy.17.1.2 Income and Income Inequality to Health
The distribution of income may have consequences for population health through two broad mechanisms. First, the health of an individual may depend on his or her (parents’) level of income. If health is a normal good, then demand for it rises with income, and the relationship should be stronger in countries that rely more on the market to allocate health resources, in particular medical care. Second, some claim that the health of an individual is contingent not only on his own income but also on the economic inequality within the society in which he lives (Wilkinson, 1996; Wilkinson and Pickett, 2010). Aggregate data show a clear negative association between measures of population health and income inequality. One proposed mechanism is that psychosocial stress arising from the stigma attached to low relative incomes is physiologically damaging. But the negative relationship between average health and income dispersion could also arise from decreasing health returns to absolute income (Gravelle, 1998; Rodgers, 1979).
We weigh the evidence that not only income but also income inequality has a causal impact on health and so affects the distribution of health in a population.17.1.3 Scope of the Chapter
The literature on the socioeconomic determinants of health is immense and comes from epidemiology, sociology, demography, and psychology, as well as economics. We confine attention to the relationship between income and health, which has been the focus of the economics discipline. Our goal is to establish what is known about the relationship from empirical analyses, and we do not cover the normative literature on health inequality. The evidence we assemble is relevant to the ethical judgment of distributions of income, health, and well-being, but we do not discuss how such normative evaluations might be conducted. Interested readers can consult the excellent discussion of some of the normative issues by Fleurbaey and Schokkaert (2011), as well as Chapters 2 and 4. Relat- edly, we do not cover the burgeoning literature on the measurement of income-related health inequality (Erreygers andVan Ourti, 2011; VanDoorslaerand Van Ourti, 2011). Inequality in both income and health could also be analyzed using measures of multidimensional inequality, which are discussed in Chapter 3.
Population health is a standard covariate in empirical growth models, and its contribution to growth has been the focus of a substantial literature aiming to estimate the economic returns from health investments (Barro, 2003, 2013; Commission on Macroeconomics and Health, 2001). We do not cover this literature on the relationship between average income and health because it says nothing about the distribution of each variable across individuals. We do cover evidence on the impact of individual health on income and of income on health, looking at low-income, as well as high-income, countries. But the balance is tilted toward a focus on the latter. Comprehensively covering the very large literature on the impact of health (and nutritional status) on earnings in low- income settings (Strauss and Thomas, 1998) would be too unwieldy.
We refer to this literature mainly to establish whether the income-health relationships observed in this setting differ from those in high-income economies that we consider in more detail.Although we have referred until now to the relationship between income and health, our scope is a little broader. We also consider the relationships between wealth and health. Wealth is an economic outcome ofintrinsic interest and is arguably a more appropriate indicator of the economic status of older individuals who provide much of the action in terms of variation in health. The health-wealth effect is likely to differ from the health-income effect. Health may affect income largely through labor market returns. This will feed through to the distribution of wealth, but, in addition, ill-health may threaten wealth through asset depletion to pay for medical and nursing care.
17.1.4 Organization of the Chapter
We begin by illustrating the strong positive relationship between health and income, using data from three countries—China, the Netherlands (NL), and the United States (USA)—that differ greatly with respect to level of development, economic inequality, labor market structures, and social welfare institutions. For each country, we show the contribution (in a purely statistical sense) that health differences make to income inequality, and, from the other side, the extent to which income variation accounts for health inequality. Having established the strength of the association between income and health, in sections 17.3 and 17.4, we turn to the mechanisms potentially responsible for the relationship and the extent to which it arises from a causal effect of health on income and vice versa. Section 17.3 identifies a number of routes through which health may impact income and wealth, paying particular attention to how economic inequality may be generated by health differences. The pathways considered are wages, work, human capital, early-life health risks, occupation, marriage, and medical expenditures. Evidence relevant to each broad pathway and more specific mechanisms is reviewed. Section 17.4 looks at the relation from the other direction: income (wealth) to health. Much of this discussion concerns whether income (wealth) has a causal impact on health over and above that of other socioeconomic characteristics, such as education and occupation, and after controlling for correlated determinants, such as time preferences and risk attitudes. Section 17.5 considers the logic and empirical support for the hypothesis that health is determined by economic inequality and by relative, as opposed to absolute, income. Finally, section 17.6 briefly summarizes the lessons that can be drawn from the literature about the nature of the income-health relationship and discusses what these imply for the normative evaluation of the distributions of income, health, and well-being.
17.2.