When Do Currency Unions Increase Trade?

A stylized empirical fact in the international trade literature is that currency unions increase trade. This “Rose effect” suggests that a country pair that shares a common currency will, on average, trade three times as much as countries that don’t. In this paper, I question whether currency unions have heterogeneous effects over the distribution of the trade variable. The motivation is that regressions reported in the previous literature give average effects, while common currencies can affect countries’ trade differently over the trade distribution. I build on the same gravity approach and dataset of Rose (2000) to allow easier comparison with existing literature and employ newly developed quantile treatment effect techniques to study what is happening at different quantiles of the trade distribution. Estimation results suggest significant amounts of heterogeneity in the effect of currency unions on bilateral trade.