Economic Status as a Determinant of Mortality Among Black and White Older Men: Does Poverty Kill?
Citation: Menchik (1993) Economic Status as a Determinant of Mortality Among Black and White Older Men: Does Poverty Kill?.
Internet Archive Scholar (search for fulltext): Economic Status as a Determinant of Mortality Among Black and White Older Men: Does Poverty Kill?
Tagged: uw-madison (RSS), wisconsin (RSS), sociology (RSS), demography (RSS), prelim (RSS), qual (RSS), WisconsinDemographyPrelimAugust2009 (RSS)
Summary
Introduction: In Britain, a large literature exists which examines the effects of social class on mortality. The results of this research have confirmed the connection, and even found evidence for a gradient effect. However, in the U.S. relatively few studies examine the relationship between income and mortality, and the studies which do exist are at a high level of aggregation, making causal inferences at the individual level tenuous. Also, it is important to distinguish transitory measures of income (e.g., annual household income) from more permanent measures of income (e.g., wealth). Finally, it is critical to include measures of family background-including ethnic group-and other environmental effects in models that examine the relationship between income and mortality. Menchik's study is designed to address these issues.
Methods: Menchik uses National Longitudinal Survey (NLS) data to test his hypothesis that income is a key determinant of mortality. The use of longitudinal data at the individual level enables him to establish causal mechanisms and also to separate the effects of stocks from flows. Permanent income was measured by (1) net household worth in 1966 and (2) average annual earnings. Transitory income was measured by an index which measures the ratio of time spent in poverty over the course of the study. Background variables include measures of parental mortality, parental education and racial identification (i.e., black/non-black). Environmental variables include a dummy variable for living in the South in 1966 (start date of the study) and for living in a small town. Additionally, variables on subject's baseline health, years of completed schooling and marital status were included. The dependent variable is whether the subject is dead by 1983. Of the 5,020 men in the NLS in 1966, 1,498 died by 1983. Analyses consisted of a series of logit models. Findings: Without the introduction of controls, Menchik finds that people in the poorest group were three times more likely to die than people in the wealthiest group. When control variables are introduced, the negative effects of both permanent income ("stocks") and transitory income (time spent in poverty) were confirmed. Moreover, given the longitudinal nature of this study, it was possible to confirm that income has a causal connection to mortality. In fact, permanent income variables alone explain roughly half of the mortality difference between blacks and whites. When poverty and other controls were introduced, about 75 percent of the black-white difference was accounted for. Moreover, permanent income greatly reduces the effects of education on mortality. When all controls are included, the independent effect of schooling disappears.
Conclusions: "The evidence presented in this paper shows that differential mortality by economic status is strongly present in the United States to-day (sic), and that this relationship is monotonic, with death rates of wealthier men being lower than those of less wealthy men. Also, the greater the number of spells of poverty, given permanent income, the higher the death rate. these data suggest that differential mortality by economic status is confused with the well-known difference in mortality by ethnic group. An implication of this paper, therefore, is that differences in mortality between ethnic groups are, in large part, a consequence of poverty or low permanent income, as opposed to genotype" (p. 436).
