What is “African growth”?

Lant Pritchett provides some numbers to underscore a classic argument:

Perhaps the best thing the developed world could do for the growth prospects of Africa is to stop talking about the growth prospects of Africa…

The growth rate of GDP per capita across 155 countries in the world from 2000-2005 (using data from the latest Human Development Report) was 2.2% per annum and the standard deviation of that growth rate was 3.8.

Among the 21 countries in Western Europe the average growth rate over this period was 3.5% and the standard deviation among countries in Western Europe was 1.5. Now that’s a pretty good aggregate, knowing that country X is in the group “Western Europe” shifts my priors a bit upward, European growth was better and reduces my uncertainty about its growth rate by a lot—I am pretty sure it didn’t have negative growth nor growth at 8%.

Now take the 45 countries in Sub-Saharan Africa. Over 2000-2005 the average growth rate was 2.2%—exactly the global average—but the standard deviation among African countries was 6.1%—much higher than the global variance. This is a terrible aggregate. All knowing that country X is “African” has done for me is increase the variance—I am not sure whether it was growing very fast (as were Sierra Leone and Mozambique) or collapsing (as were Liberia and Cote d’Ivoire).