It is the conventional wisdom that urban centers with their concentrations of human and physical capital and their dense social networks are engines of growth. One exception to this is can be the case where a dominant urban center by virtue of its institutional monopoly on a country or region’s economic life is able to extract economic rent from the surrounding country side and ultimately kill the goose that laid the golden egg.
The potential relationships are interesting ones. Do countries with a large share of their population residing in the capital city create an environment of economic rent seeking that siphons resources from the rest of the country into the capital via public spending and regulations that ultimately weaken aggregate economic performance? This is probably the question for a thesis in urban economics but why not have some fun with it. I took per capita GDP data in U.S. PPP$ for 30 OECD countries in 2007 along with an estimate of their population and used the 2008 World Almanac to find the population of each country’s capital. I calculated the share of population accounted for by the capital city and plotted it against the level of per capita GDP
The result? Well, there really was not much of any pattern. There were two obvious outliers –Iceland and Luxembourg – so I removed them – and plotted again with a trend line. The result in Figure 1 could not find a negative relationship between per capita GDP and the share of population residing in the national capital.
Does this mean there is no relationship? Well, this is pretty crude empirical analysis. A richer data set and some control variables would be helpful. The data set probably also needs to be expanded beyond the set of OECD countries, which are by and large relatively prosperous and stable countries. However, one final experiment. Canada’s provinces vary quite a bit in terms of the provincial population share accounted for by the capital city from a high of 59 percent in Manitoba to a low of 8 percent in British Columbia (Figure 2). When the capital city population shares are plotted against per capita GDP for the year 2006 you get what could be interpreted as a hump-shape – per capita GDP grows as the population share grows up to about 30 percent but then increases in the share are associated with a decline in per capita GDP (Figure 3).
This might suggest that a provincial capital that grows as a share of the province’s population has a positive impact on per capita GDP growth but once it reaches a critical point of about 30 percent of the population it starts to exert a negative effect. Perhaps once a third of the population lives in the capital, the critical mass exerts political dominance on provincial life out of proportion to its population size. The Canadian provinces that fall into this boat would be Newfoundland & Labrador, Nova Scotia, Ontario, Prince Edward Island and Manitoba. However, there are only 10 data points here and only for one year. Get rid of Alberta (the peak point) and you are back down to a flat line relationship between the two variables. However, this might be an interesting topic for a graduate student to tackle.
In terms of a specific application, I am particularly curious if this type of story might explain why in Ontario over the last decade so many elementary and secondary schools have been closed outside of the Toronto area while the Toronto schools have not only managed to keep more small schools open but also maintained funding for things like swimming pools and teacher-librarians. Apparently, 19 percent of schools in Eastern Ontario and 10 percent of schools in northern Ontario have teacher librarians compared to 92 percent in the GTA schools. Did close proximity to the provincial decision makers create opportunities for advocacy not available outside the capital?