I've worried before about the applicability of the Growth Diagnostics approach to the poorest countries, especially those in Africa.
Well Dani Rodrik has just published an updated list of studies that he is aware of.
Of the 19 studies, only 2 (Kenya and Uganda) are on Sub-Saharan Africa (excl. middle-income South Africa).
Seven (Afghanistan, Bolivia, Cambodia, Kenya, Moldova, Mongolia, Uganda) are on Paul Collier's list of "Bottom Billion" countries (hold on - Moldova?! isn't that in Europe!?).
Is this a data availability issue? Do African countries have atypically poor data, or is just that they are so poor?