The latest Economist has an interesting feature on inequalities among regions within countries. The article compares countries on their ranges in GDP per head (the ratio of richest region to poorest). Thus, we get charts like the following:
But range is an extremely crude concept for measuring inequality. In the U.S., the District of Columbia is by far the “richest” “state,” because its large number of commuter workers generate large GDP without figuring into the denominator. Moreover, the use of the range to illustrate dynamics over time is misleading:
This chart makes it appear that the U.S. has rapidly growing regional inequalities. But the increase here is being driven by D.C. again. The growth of the federal government has concentrated ever more GDP in the District, causing its numbers to look increasingly out of whack with the rest of the country.
A better approach is to compare rates of regional GDP per head convergence. Convergence is the phenomenon whereby poorer economies tend to “catch up” to richer ones. A rough-and-ready benchmark for “good” convergence is an annual rate of about 2%. Econometricians derive rates of convergence in GDP per capita by regressing annualized GDP per capita growth on initial GDP per capita for a dataset of economies. I have calculated regional convergence rates for Canada (provinces and territories), the U.S. (states and D.C.), and the European Union (member states before 2006) over various periods. Here are the results:
The “equalization” column indicates whether the federal system has extensive equalization payments that give grants to poorer regions. The EU does have a nominal equalization program, but it does not redistribute much money. Of these systems, only Canada has a truly extensive equalization program.
Despite this, Canada’s convergence record is the worst of these systems, although the differences between the U.S. and Canada are small. Over the entire 1981-2005 period, U.S. states converged at 1.9% per year, while Canadian provinces did so at 1.6% per year. The EU clearly has the best convergence record, with a massive 8.0% annual convergence rate during the 1995-2005 period, which saw the rapid rise of Ireland, Greece, Spain, and Portugal, relative to the rest of the EU. (Eastern European countries are not included in these numbers, because they had not joined the EU yet.)
This evidence suggests that decentralized federal systems do a pretty good job of getting rid of regional inequalities, even without equalization programs. In a paper currently under “revise-and-resubmit” at an economic geography journal, I present much more formal and systematic evidence to this effect. If and when it is published, I will revisit the topic.