Much economic research is limited in application, perhaps relating to very special country circumstances and behaviour of people. Occasionally, a piece of research is far broader in its application, even though the superficial level of sophistication is no different from the more limited papers.
One of these disproportionately useful theories describes optimal currency areas. The essence of the theory was outlined in a single, short paper in 1961. The theory claims that the geographical areas most suitable for a single currency will have the following properties:
- the areas have trade integration
- if one area has a net trade surplus with another area, it is offset by a net capital deficit
- they have similar responses to external economic events
- they have similar macroeconomic character, in terms of quantities like inflation and government expenditure
- there is free movement of goods, people, and money between them.
The rough idea is that if these conditions hold, then not too many tensions will emerge between the areas under a single currency, and any tensions can be resolved by market mechanisms. Recent theories have refined the ideas, making allowance for things like sticky prices and capital market instability. But the core of the theory has proved remarkably robust, and is recognisable in the criteria which trading blocs such as the European Union apply when assessing whether they should have a single currency.
I suspect the theory has been so successful because it condenses and reapplies some other, well-respected parts of economics like the theory of Pareto optimality to produce a composite which is so logical that there is no need for extended exposition or analysis. In some ways, such an approach is dangerous - one would like an extensive empirical analysis of whether optimality is indeed a consequence of the criteria, and an analysis of whether alternative definitions of optimality could apply, and a full analysis of conditions for failure. The theory of optimal currency areas is so well engineered, however, that is seems likely that it would still be a theoretical core of any new theory, if only to reject the optimality of areas which deviate widely from the conditions.