Feeling the Squeeze: The Effect of Birth Spacing on Infant and Child Mortality During the Demographic Transition

Abstract

A negative association between birth spacing and infant and child mortality has been consistently identified within modern populations in developing countries. Generally speaking, children born following shorter birth intervals have been found to have higher mortality (Hobcraft, McDonald, & Rutstein, 1985; Kozuki et al., 2013; Rutstein, 2005). The reasons for this association are unclear, however. Leading hypotheses attempt to explain these differences as a result of sibling competition, maternal depletion, infection transmission, or unobserved maternal factors that simultaneously influence fertility and infant mortality (e.g. inadequate breastfeeding practices), but none has had overwhelming support. This study contributes to this body of research in a few important ways. First, it examines this association in a historical context, which has largely been ignored, and documents how the relationship evolved over the course of economic development. The data come from the Roteman Database, a longitudinal register kept for Stockholm, Sweden between 1878 and 1926. Second, it attempts to isolate some of the hypothesized causal mechanisms by studying variation within families by controlling for maternal fixed-effects thereby eliminating the potential for compositional differences among mothers to drive this relationship. Third, it exploits variation in the timing of death of older siblings to assess the plausibility of the different causal mechanisms. Results suggest that the relationship between preceding interval length and mortality holds even when accounting for unobservable maternal factors, particularly for post-neonatal and early childhood (age 1-4) mortality. The association disappeared over time as infectious disease mortality fell and deaths of previous siblings during the postnatal period disproportionately tended to increase the risk of dying among index children. These findings strongly suggest that the relationship is related to the transmission of disease between siblings.