Do Currency Unions Grow Too Large For Their Own Good?

Abstract

This article puts a rigorous foundation under the proposition that currency areas, as they admit more members, face a rising marginal cost curve which cuts the marginal benefit curve from below. However, at any given time, the median member faces lower marginal cost than the average member, so that, if new members are admitted by majority vote, and existing members are myopic, the currency area will expand beyond its optimum size. Although the currency area imposes negative externalities on countries outside it, we find that the existence of one currency union has no effect on the costs or benefits of forming or enlarging another. An optimum currency area, if it means anything, must mean an area where the last recruit conferred more benefits than costs on existing members, but the next one will do the reverse. The difficulty of showing whether such a case could exist in theory, let alone pointing to real-life examples, has meant that most 'optimum currency area theory' over the last 35 years has, instead, concentrated on the easier question of what kind of partners you should look for if you are contemplating forming, joining or extending a currency union (Mundell, 1961; Goodhart, 1989; Kenen, 1969). Even papers which do consider the existence of an optimum have seldom made a rigorous effort to show that the MC curve (as more members are admitted) cuts the MB curve, if at all,