Exactly what are African growth statistics based on?

Not much that is good, according to this paper:

African economic development and economic history has been dominated by scholars more concerned with the quantity thanthe quality of data. Any data are of course said to be better than none, but beyond a certain point this ceases to be true. Despite well-known quality issues econometricians use all data available to examine economic progress.

Jeffrey Williamson, a leading economic historian using quantitative methods, once rhetorically asked: ‘Have you ever met a cliometrician who throws data away?’

That is Morten Jerven, who investigates how politics and context and history matter in the production of economic data. The continent’s ‘lost decades’ are lost in more than one sense.