Currency Unions and Trade: How Large is the Treatment Effect?∗

It has been conspicuously hard to empirically identify clear-cut effects of Þxed exchange rates and other monetary regimes on the real economy. This paucity of empirical Þndings has lead many academic economists as well as other observers to the view that radical monetary reform — such as the adoption of a common currency — might have limited economic consequences. For example, it probably explains the common view that the EMU is “Þrst and foremost a political rather than an economic project”. A recent Economic Policy paper by Andy Rose challenges the conventional wisdom (Rose, 2000). He uses evidence from existing currency unions in the world economy to estimate the effects of a common currency on trade. According to his regression estimates, a currency union expands bilateral trade between two average member countries by a dazzling 200 percent or more. Given the novelty of the Þnding and the importance of the issues, these results have received considerable attention. Critics voiced a number of concerns about Rose’s methodology, questioning the accuracy of his Þndings. Provoked by the critique, Rose (2000) conducted a large battery of robustness checks, however, showing that the central result holds up to the points raised by the critics. Recent work by Rose and van