Extraordinary African cities? Zipf’s law and the emerging African system of cities☆

• Strong evidence is found of agglomeration economy effects driving the African city size distribution ever closer to follow Zipf’s power law. 1 1 “African cities” refers to Sub-Saharan Africa. • Africa is evolving towards a common spatial urban equilibrium which in the long-run is similar – rather than dissimilar – to that in OECD countries. • Both dispersed urbanisation and excessive concentration of economic activity are found to be detrimental to Africa’s economic development. • The impact of Zipf’s power law on economic growth depends on a country’s colonial legacy, level of income, and land size. • Armed conflicts skew urbanisation process away from a country’s spatial equilibrium, further damaging its long-term prosperity. This paper revisits the rank-size power law of urban hierarchies in Africa, with new analysis of the relationship between urban agglomeration, colonial legacies, and spatial economic inequalities. It explores urban patterns of spatial distribution of economic activity across Sub-Saharan Africa and finds evidence that Zipf’s law holds for African cities. Based on population panel data from Africapolis for 34 countries over the period 1960–2015 – it finds that the rank-size rule relationship has strengthened over time, with the statistically significant parameters converging closer to −1. The empirical evidence supports an inverted U-shaped relationship between the Zipf parameter and GDP per capita growth with different effects across groups of countries by colonial legacies (British versus French), income, and land size. These results are based on 2SLS and dynamic system GMM panel data quadratic estimations. The findings suggest an urban Kuznets bell-shaped curve, implying that economic growth may initially increase urban income inequalities, while reversing to reduce spatial income inequality with further GDP per capita growth. The African urban hierarchy – both nationally and region-wide – is evolving to the empirical regularities commonly observed elsewhere.