« How not to evaluate immigration policy | Main | Should economists be licensed? »


Feed You can follow this conversation by subscribing to the comment feed for this post.

The % in the capital city could be low for two reasons. 1. the province/country has a rural population/agricultural economic base or 2. the capital is located in the province's/country's second or third city (e.g. Victoria, Regina,...). It would be interesting to try to differentiate between these two different explanations.

I think your analysis of figure 3 is incorrect. Surely the "hump shape" is defined by Alberta's oil wealth and therefore not indicative of a trend. Take out Alberta and your scatterplot is trend-less. Because Newfoundland, Ontario NS and PEI have the same fraction in the provincial capital they can define am "error bar" (ie a 1-sigma deviation) and by eye it looks like a few thousand dollars. With that kind of fuzz I don't see your conclusion.



Thanks for pointing out that Alberta is the peak. I had attributed the peak to Newfoundland and Labrador which I have now corrected in the text of the post.
As for your observation that without the peak, the diagram is trendless, well, I know. I'm trying to be provocative and am fully aware that 10 data points is inadequate and in fact I say: "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." I think a rigorous empirical assessment would require a fairly substantial pooled-time series cross-section going back at least half a century if not longer.
Differentiating the provinces that have a strong second city from those that do not is a good point. Saskatchewan and Alberta, for example, have strong second cities while Manitoba in particular and even British Columbia do not. Ontario is dominated by Toronto but also has the federal capital region and that needs to be considered. More data definitely required.

Livio, as you say, it seems you'd need to look at a weaker state or one subject to popular unrest to see if this relationship holds. I think Afghanistan, Pakistan or India might be interesting countries to analyze, for various reasons.

In Canada, the effect of proximity to the people in large capital cities might be offset by the under-representation of urban areas in our electoral system. One riding in Brampton, for instance, has nearly 200,000 residents, while the average is around 110,000.

At least in Canada I don't think the causation would run Big Population in Cities -> Rent Seeking behavior. The relationship probably holds better in dictatorships or countries with weak governments.

For sure, check out Portugal. A unitary state with strong central government centred in the capital city Lisbon. If that isn't a poster child for rent seeking activities by urban residents of the capital versus the rest of the population I don't know what else is.

The claim to fame of the next largest city of Porto in northern Portugal is for its hard working and industrious citizens. That must tell you something.

CBBB: Look into the fairly recent book, "Soccernomics", which touches on the relationship between economic rents and capital cities from a soccer perspective. Specifically, the past successes in the capital cities of (former) dictatorial countries (i.e. Spain, Portugal, Yugoslavia) and democracies (i.e. England, France, Germany) tend to differ, with the second/third cities performing much better in the democracies.

I've long had an interest in economic geography, economic history and urbanism. I cannot think of references off-hand, but I do recall that as countries urbanize and industrialize, the major city at first grows a lot more rapidly that other cities, but then in the later phases, the growth of the largest city slows down and secondary cities grow much faster.

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad