Contrary to the image of African countries as static mono-exporters, it is unpredictable from one period to the next which will be the top exports in each country. Consider this picture of Tanzania’s top exports in 1998 and 2007.
This is pattern of rapidly changing success is the norm across African countries. If you take the top 100 exports in each country in 1998 (or the first year in which data is available), its correlation with the rank of those same exports in 2008 is only .29.
Moreover, almost none of the changing success is explained by global commodity prices. In fact, there is little difference in the dynamic changeability of African commodity export performance and that of the continent’s non-commodity export performance. Nor is there any difference between how much global prices explain commodity exports (which is hardly at all) compared to non-commodity exports.
The usual stereotype of African exports as just given by a natural commodity or mineral endowment, with fluctuations mainly explained by global commodity prices, is just ... wrong.
These findings were featured in a paper by Ariell Reshef (UVa) and myself in the National Bureau of Economic Research conference on African Development Success July 18-20 in Accra.
What does it all mean? Actually, the patterns in Africa were similar to those in non-African countries. In all cases, succeeding in exports requires aiming at a moving target. Who will do better under these conditions, state industrial policy planners or decentralized entrepreneurs with specialized knowledge of what is working and what is not in each sector?