Are South Africa’s cities too small?

Abstract

Are South African cities to small? Given the history of South Africa’s spatial development, one might expect that South African cities might be under-sized, and not over-sized as in many other developing countries. It is found that the rank-size distribution explains the sizes of South Africa’s cities but that Zipf’s Law does not hold for the country’s cities. The so-called q-coefficient was found to be equal to −0.75 for the 123 places with population in excess of 100 000. It was also found that urbanisation in South Africa over the past decade seems to have taken the form of the parallel (slow, 1.04%) growth of five large cities. Finally, calculating the “ H-measure” for 19 metropolitan areas in South Africa yields an inverse H-measure of 11.3. This suggests a reasonable degree of dispersal, which would only be consistent with optimal city size if transport costs were low and manufacturing not in need of scale economies; two conditions unlikely to apply to South Africa. Finally, the primacy ratio for South Africa’s largest urban agglomeration was found to be 38%. This suggests that the size of the Johannesburg-East Rand urban agglomeration (the primate city) may be relatively too large, whereas more efficient growth may come from larger harbour cities.