Sub-Saharan growth surprises: Geography, institutions and history in an all African data panel

We use two types of cross-country growth regression models to revisit explanations of slow growth in Africa looking at growth rate variation among African countries only. Both sets of models produce results that are surprising given conclusions based on global sample: within Africa, we .nd greater coastal population negatively and greater ethnic heterogeneity positively associated with growth, while distance from the equator is at .rst negatively and only later positively associated with growth. Our results suggest also that institutional and policy variables are endogenous to geographic and historical factors including the colonizing power and the religious and ethnic make-up of the country.