Differences in factors such as income, education and marital status could contribute overwhelmingly to the gap in life expectancy between blacks and whites in the United States, according to one of the first studies to put a number on how much of the divide can be attributed to disparities in socioeconomic characteristics.
A Princeton University study recently published in the journal Demography reveals that socioeconomic differences can account for 80 percent of the life-expectancy divide between black and white men, and for 70 percent of the imbalance between black and white women.
Numerous existing studies on the topic have found that mortality differences are associated with certain socioeconomic disparities, but have not determined to what extent the life expectancy gap can be explained by such contrasts, noted author Michael Geruso, a doctoral student in Princeton's Department of Economics.
Geruso pulled mortality information from the National Longitudinal Mortality Study (NLMS), a nationally representative survey of households led by the U.S. Census Bureau that examines the demographic and socioeconomic factors related to death. He concluded from this data that the average life expectancy from age 1 - the NLMS does not fully capture neonatal and infant mortality - as 71 for white men and 66 for black men over the study period. White and black women lived an average of 78 and 74 years from age 1, respectively. These figures match statistics from the U.S. Centers for Disease Control and Prevention.
Geruso then focused on racial divergences in seven areas: family income, education, occupation, unemployment, urban residence, home ownership and marital status. He determined the percentage each variable contributed individually to racial differences in life expectancy, and also examined the influence of various combinations of factors.
As a stand-alone factor, income accounted for 52 percent of the life-expectancy difference between black and white men, and 59 percent of that for women. For both genders, income was twice as high as the second most significant factor, education. After controlling for income and education, racial imbalances in other factors exhibited little power to explain racial differences in mortality.
Geruso also found that the influence of some factors depends on a person's gender, some of which were surprising, he said. Notably, marital status could account for more than 10 percent of the life-expectancy gap between white and black men, but seemed to have a negligible influence on mortality differences between white and black women.
Geruso explains his findings as follows:
"The most important contribution of this research is that it evaluates the relationship between socioeconomic status and racial gaps in mortality for almost all ages and using an unusually rich set of socioeconomic variables. It offers a precise, quantitative answer to a question many people are interested in: How much of the glaring disparities in life expectancy between blacks and whites might be explained by differences in a relatively small set of socioeconomic characteristics?
"What was really shocking to me was that no one had already done this study. The basic research question is one that has certainly been asked over and over, and dozens of good studies have chipped away at it. But for a variety of reasons, past studies haven't offered a precise or complete answer, either because of a lack of detailed data or methodological limitations.
"Instead, many existing studies are limited to examining mortality differences over a narrow age range and using incomplete data on socioeconomic variables - for instance, income but not education, or vice versa. The interested reader could perhaps cobble together the results of a dozen studies, but that would provide an impression, not a result. One would be left with only a qualitative sense of how the relationship between socioeconomic status and mortality might contribute to the life expectancy gap between blacks and whites in the United States.
"My basic strategy was to 'reweight' the sample of black respondents in the NLMS data so that the distribution of all socioeconomic variables among blacks matched that of whites. For instance, when calculating life expectancy I scaled down the contribution of low-income black respondents and scaled up the contribution of high-income black respondents. That way, the effective proportion of low-income people in the black and white samples was even. This told me to what extent the life expectancy gap closed between the race groups once income was equalized.