When Birth Spacing Does and Does Not Matter for Child Survival: An International Comparison Using the DHS

Kieron Barclay, Max Planck Institute for Demographic Research
Martin Kolk, Stockholm University
Joseph Molitoris, Hungarian Demographic Research Institute

A large body of research has shown that shorter and longer birth intervals are associated with higher risks of child mortality, leading to policy recommendations to space births 3-5 years apart. There is little agreement, however, about whether or not this is a causal relationship, and what causal mechanisms facilitate it due to two shortcomings in the current literature on the subject: a failure to control for unobserved heterogeneity that may lead to biased estimates and a lack of international comparisons to understand variation in the relationship. In fact, the former criticism has been recently used to dismiss the notion that birth intervals influence mortality risks at all (Klebanoff, 2017). We address both of these by conducting a within-family survival analysis of over two million children from over 80 countries at various levels of development. We find that shorter birth intervals do indeed cause higher child mortality risks after controlling for unobserved heterogeneity, but that longer intervals do not. This comes with a caveat, however. The protective effect of longer birth intervals is exclusively relegated to low levels of development. Our findings show that the large body of work on the subject cannot be outright dismissed, but also that a relationship between birth intervals and child mortality risks should not be expected across all populations. It is primarily the worst off for whom birth intervals matter.

Presented in Session 1169: Fertility