Estimating Neonatal Mortality Rates from the Heights of Children: The Case of American Slaves

Underenumeration of vital events is a problem familiar topeople who work with historical demographic records. This paper proposes a method for recovering information about neonatal mortality.The approach utilizes average heights of young children to predict the birth weight of American slaves. The results suggest that slave newborns weighed on average about 5.1 pounds, which places them among the poorest populations of developing countries in the mid-twentieth century. The birth weight distribution and a schedule of mortality by birth weight suggest that previous estimates of slave infant mortality are too low. The poor health and stature of children and the relatively large size of slave adults is a pattern of growth and development that is unobserved among poor populations of the twentieth century. Thus slavery may have created an unusual pattern of nutritional resource allocation across ages.