Cities and firms

In an interesting overview of economic geography and the Sub-Saharan African economy, Tony Venables cites a paper looking at productivity and the location of firms in cities:

The quantitative evidence of the productivity effect of city size comes largely from studies of cities in developed countries.Rosenthal and Strange (2004) report a consensus view that doubling city size is associated with a productivity increase of some 3 – 8%. This is a large effect – moving from a city of 100,000 workers to one of 3 million is predicted to increase productivity by more than 30%. Au and Henderson (2006) find even larger results for Chinese cities, where they estimate that moving from a city of 100,000 workers to one of 1.3 million workers raises productivity by 80%, although beyond this scale weak diminishing returns cut in.

Looking at this for the African context and running a regression equation for city size on independent variables including , he finds that both country population and country area are highly significant determinants of city size.  Further:

a merger of two similar size countries – i.e. a doubling of population and area – would lead to a 75% increase in the size of the largest city.... which suggests that the smaller size of African cities is due, in large part, to the fragmentation of countries.

The punchline for me is this:

Africa’s fragmentation and consequent urban structure may have impeded the development of major international manufacturing centres of the type that contribute to the performance of high growth economies. 

I'm particularly interested in this issue on the back of having read Jane Jacobs' Cities and the Wealth of Nations. Jacobs spends a long time discussing how cities can integrate entire regions in a dynamic process of wealth creation through replacement of goods previously imported to that city (not necessarily to the country)  or else stagnate due to a lack of that process of import replacement and a decline in the degree of integration with the surrounding area. She talks of city-regions, the areas around cities which come under its economic influence and how these only develop for certain cities.

It would be very interesting to relate this kind of analysis to enterprise data on firm productivity and characteristics within different cities, and especially to the kinds of results we find moving from city to city in sub-Saharan Africa.....

One day when I get that post-doc...

(Actually, the World Development Report 2009 discussed all this stuff so I will be looking in there for more insights...)