If you refounded North America, how many currencies would you go for? Whatever the answer, you probably wouldn't insist on Vancuver and Seattle being different.
The whole thing is reminicisent of the debate around Europe. In May, a JP Morgan research note revealed that the Eurozone was more diverse than pretty much every other possible monetary union:
The x-axis is a measure of similarity between countries. It measures over 100 economic, social and political characteristics. Michael Cembalest, the report's author, then applied this measure to 11 hypothetical monetary unions, as well as to the major countries of the Eurozone (he excluded smaller countries like Cyprus and Malta, but the results aren't that different if they are included; nor does the inclusion of Greece affect the results all that much).
What he finds is that many monetary unions that came close to existing exhibit far more similarity than the Eurozone. This includes Latin America, the Gulf states, and Central America. He then pushed it further: reconstituting several former empires, including the USSR, Ottoman Empire, and the British Empire in Africa, would also result in unions with more similarity than the EU.
Now, three academics from the Democritus University of Thrace have performed a similar analysis on the US and Canada, and found that – economically, at least – the present borders make little sense. E. Chrysanthidou, P. Gogas, and T. Papadimitrioy apply Robert Mundell's theory of Optimal Currency Areas (OCA) to the hypothetical issue of a north American currency union.
An OCA is an area where the macroeconomic conditions between two or more regions are suitable for creating a monetary union. All such unions have potential benefits – eliminating currency risk means that conditions are much more favourable for trade within the union – but they also have potential downsides, as the eurozone is demonstrating presently. If the various involved regions are similar enough, the benefits are likely to outweigh the risks.
The theory, which stems from the 1960s, was originally based on an examination by Mundell of the US and Canada, but it took on a more practical bent with proposition of the European Monetary Area. Since then, it has been largely applied to Europe and similar cases of actually-existing, or at least widely proposed, currency unions.
The authors return to the source, and attempt to work out, using two different methods (Correspondence Analysis and Hierarchical Cluster Analysis), what the groupings between the fifty US states and ten Canadian provinces ought to be.
The conclusion is not two, but three different countries, one on each coast and one in the middle:
The authors describe the differences:
The first one includes regions mainly from the East that are industrialized, and characterized by high levels of economic activity as this is measured by the macroeconomic variables used in our analysis.
The second part includes regions mainly from western US and Canada with diverse levels of economic activity and prosperity.
Finally, a third group of regions can be identified. This group includes a geographically diverse set of regions as it spans from east to west. The common factor though that links these regions is the relatively low level of economic prosperity as it is measured in our study in terms of income, growth, imports, exports, etc.
It would be rather awkward, to be sure – but no less awkward than the current arbitrary line drawn along the 49th parallel.