The Double African Paradox: What does selective mortality tell us?

We study the relationship between height stature and child mortality in West Africa. This is motivated by two things: understanding the determinants of height, widely used health indicator, and explaining the « double African paradox ». This paradox comes from the fact that Africans are relatively tall in spite of extremely unfavorable income and disease environments; this the level paradox. The second paradox is that African height stature decreased in recent years, despite better health conditions and lower child mortality; this is the trend paradox. These stylized facts are surprising as both child mortality and height stature are viewed as health indicators, so that we expect a negative correlation between the two. To study this paradox, we focus on West African countries only, where child mortality levels are very high. For Benin, Burkina Faso, Côte d'Ivoire, Ghana, Mali, Nigeria and Senegal, we use DHS (Demographic and Health Surveys) data to measure child mortality (before 5) at the region X period level using the retrospective birth history of mothers. We want to test to what extent the paradox can be explained by selective child mortality. More generally, there is a need to understand how much child mortality levels and trends affect the study of height stature in Africa. Instrumentation can not be used in our context as we would need an event that increases or decreases mortality without affecting nutrition and regardless the distribution of heights. Consequently, we build a statistical model that we estimate linearly and nonlinearly. We first show that the correlation between adult height and mortality within region is not significantly negative in our setting. By estimating a nonlinear relationship between height and child mortality, we show that in high mortality contexts, when child mortality decreases selective mortality decreases as well, so that more short people survive. We are able to explain some differences in height levels between African countries and countries where selective mortality is lower. We also manage to explain why height stature did not increase as much as expected in Africa compared to the decrease in child mortality rates in the second half of the 20th century.