A recent paper in Nature Energy has grabbed headlines such as ‘Africa’s green energy transition unlikely this decade’ on the BBC. The paper by Oxford’s Galina Alova, Philipp Trotter and Alex Money applies a project-probability model to a database of nearly 2500 planned power projects and concludes that Africa is on a path to locking in a fossil fuel-dominated future. To achieve a low-carbon transition, the authors call for a “rapid decarbonization shock” including “the exit of [development finance institutions], both national and international, from fossil fuel projects.” Understanding the project pipeline in Africa is indeed important because the continent’s infrastructure needs are so tremendous and large projects have very long lead times and, once built, operate for decades. But two big issues throw serious doubt on the paper’s policy conclusion.

First, the underlying data appears to vastly overestimate capacity growth in power from fossil fuels. The Southern African subregion for instance is projected to add ~22 GW of new coal capacity by 2030 (based on ~28 GW of planned coal projects with ~6 GW projected to fail). Yet data from the Global Energy Monitor, a highly credible public source, suggests more like ~11 GW for 25 coal projects across all of Africa have been announced, pre-permitted, or permitted (excluding refurbishments of existing plants or projects already shelved). Further, a closer project-by-project analysis of progress of those potential 25 projects estimates that new capacity likely to come online is less than 1 GW. Another ~3 GW is estimated as possible but very unlikely. That suggests the paper over-projects likely new coal by at least a factor of 7 and probably much more.

Projections for gas similarly appear overly aggressive. The paper forecasts ~39 GW of new gas-fired power capacity in West Africa, which would imply a more than three-fold increase over current capacity in the region in just the next 8 years. Assuming 300 MW per plant, that’s 130 new gas plants in West Africa alone by 2030. Anyone closely watching the region’s power market would likely recognize that this scale does not seem plausible. For comparison, the US Government’s Power Africa initiative identified less than 8 GW of gas-fired capacity in West Africa that they could even potentially help to bring to financial close (not operation) by 2030. The USG of course isn’t involved in all projects, but I strongly doubt they are missing 30 GW of gas-fired power that will soon be operational.

The discrepancies may partly be a question of different datasets, but, unfortunately, the data used in the Nature Energy paper is hidden behind a paywall, so we can’t see anything more specific than regional aggregates. (We had to count pixels in Supplementary Figure 7 to get the subregional estimates used above.) And part of the coal projection discrepancy may be status for ~9 GW from the Medupi and Kusile stations in South Africa (our analysis counted these as already operational). But the sheer scale of the differences suggests caution before expecting an explosion in new coal- or gas-fired power across the continent. (A friendly longbet, perhaps?)

Second, the paper could be easily misinterpreted because of the way it reports aggregates for “Africa” (and for the subregions in the supplementary document). Regional totals obscure important differences among markets, in part because a few very large markets dominate. Two thirds of the 244 GW of current generation capacity are in South Africa plus the six North African countries. This leaves just 81 GW for the remaining 48 countries where one billion people live. The distortionary effect is even more pronounced for fossil fuels: South Africa alone accounts for 86% of the continent’s total coal-fired capacity, while North Africa is atypically gas-dependent. Aggregation thus has the practical effect of making “Africa” appear far more fossil-dependent than it is. (The EIA made a similar mistake in their modeling.) In fact, the electricity fuel mix of many countries — including Kenya, Ethiopia, Malawi, Mali, Mozambique, and Uganda — are already majority renewables.

These two problems of implausible scale and misleading aggregation suggest that the authors’ call for a continent-wide ban on DFI finance for all fossil fuels is misplaced. Africa is not a country. Concerns about North African gas or South African coal should not be used to justify a blanket policy affecting the other 48 countries, all of which are negligible CO2 emitters. Just as we would not consider it fair to weigh Russia’s emissions for determining Portugal’s energy choices, South Africa’s coal-dependency should not constrain the energy options for Kenya, Mali, Senegal, or Liberia.

Moreover, the implication that those 48 African nations could somehow blow the global carbon budget is wrong. Their emissions are so low that if we took the extreme case of tripling their electricity consumption overnight using only gas, the additional CO2 would equal less than 1% of the global total. And in many of these same African countries, gas is already enabling the integration of more wind and solar and a transition away from more carbon-intensive liquid fuels for power or traditional biomass for cooking. This suggests that rather than issuing a blanket financing ban for an entire continent, development institutions should be expanding support to these low-emitting energy-hungry countries and providing maximum flexibility to end energy poverty.