As I have discussed elsewhere (here, here and here), in 2010, Ghana changed its base year for GDP calculation from 1993 to 2006, and the cumulative effect of change in methods, base year and adding new data on economic activities meant that total GDP almost doubled.
Many questions emerged following these news. Shanta Devarajan declared a Statistical Tragedy in Africa ,while others insisted that there was a Statistical Renaissance in Africa, especially here and also here. I do not see the reason to draw a sweeping conclusion on this – but rather emphasize that statistical capacity, and therefore the quality of the development indicators, is very very uneven across time and space in Africa.
One of the key questions for me has been how this affects the history of economic growth in Ghana. I have also been in touch with, and had requests from and conversations with United Nations Statistics Division, Gapminder and the Data Group at the World Bank on how to reconcile these different emerging pictures of Ghana’s recent economic past.
Now it turns out that until the series was revised to a 2006 base year the World Bank was using a 1975 base year – while Ghana Statistical Services used their own 1993 base year series until the 2006 series was published in 2010. The problem is that the new series only goes back to 2006. For illustration, here is how Ghana looked according to the series published in World Bank Data until 2012 (red) and from 2013 (blue). This is in current prices, but shows how the revision radically changed the post 2006 growth story in Ghana.
Note that there is a perfect overlap between the two series from 2005 and back to 1960.