Executive Summary: “The Income-Health Relationship and the Role of Relative Deprivation” by Christine Eibner and William N. Evans

A large body of literature that spans many disciplines has established that individuals with lower incomes have poorer health outcomes and higher mortality rates. In contrast, there is much weaker evidence of a link between income and health in aggregate data for developed countries. In many cases, cross-country differences in income are only weakly correlated or even negatively correlated with mortality. This paper is an attempt to explain these two seemingly contradictory results by examining the impact of relative deprivation on health.

A number of authors have suggested that individuals are adversely affected when they perceive themselves to be economically deprived relative to their peers. Although even low-earning college graduates may have a high income in an absolute sense, they may still feel deprived relative to higher-income peers. Authors have argued that relative deprivation influences health primarily through psychosocial stress. Individuals who feel they are economically disadvantaged compared to their peers may be depressed and disgruntled, conditions that affect health both directly (via heart disease, high blood pressure, and suicide) and indirectly (via increased smoking, poor eating habits, and alcohol abuse).

In a previous paper by us (Eibner and Evans, forthcoming, Journal of Human Resources) we demonstrated that mortality rates are higher for individuals who have higher relative deprivation compared to members of their reference group. Using restricted-use micro-level data from the National Health Interview Survey Multiple Cause of Death Files (NHIS/MCOD), we constructed a measure of relative deprivation based on work by Yitzaki that takes into consideration not only a respondent’s percentile rank in income within reference group but also the average incomes of people who make more than the respondent. We define reference groups using a combination of characteristics including state of residence, age, race, and education. Our results indicate that, even after controlling for reference group effects and individual income, relative deprivation is positively associated with an individual’s probability of death.

This paper summarizes the results of our earlier work linking relative deprivation to poor health and extends the analysis in a number of important ways. First, we demonstrate the persistence of the result by performing a number of specification tests. We examine the relationship between relative deprivation and health using a variety of definitions of health and health habits, for several different sub-samples, and using a number of different definitions of reference group. The impact of relative deprivation is statistically precise and quantitatively important across a variety of specifications.

We then examine whether the impact of relative deprivation on health can explain some of the basic relationship between income and mortality in individual-level data sets. When an individual's income rises, that person’s relative deprivation falls. Our models indicate that including relative deprivation in a standard mortality/income model reduces the coefficient on income by 50 percent, suggesting that half of the impact of rising income on mortality is due to a reduction in relative deprivation. Interestingly, models with only an income term and those with income and relative deprivation terms predict roughly the same change in mortality for a given change in income for a particular person.

Finally, we examine whether relative deprivation's impact on health can explain the disparity in results between studies with individual-level and aggregate data. To do this, we perform a simple simulation exercise where we increase the incomes of all people by 10 percent. Although each person is now richer in an absolute sense, most peoples' incomes have fallen relative to others in their reference group, generating an increase in relative deprivation. Subsequently, some of the pure income effect that should reduce mortality is mitigated by the relative deprivation effect—which increases mortality. Our simulations suggest that with a uniform increase in income for all people, the relative deprivation effect will dominate the pure income effect and aggregate mortality will increase slightly. These results therefore suggest that the negative impact of relative deprivation on health may explain the weak relationship between income and health at the aggregate level.