This paper studies the determinants of currency union membership. Geographical distance, colonial heritage, language, sizes and bilateral trade between two countries as predictors of their propensity to adopt a common currency are accounted for. To deal with endogeneity, two-step probit estimation method is performed. The estimation results show that geography, colonial heritage, size, and speaking the same language predict monetary unions quite well. However, bilateral trade does not enter significantly in the second-stage estimation, thus revealing that bilateral transactions between two countries are not a useful indicator of their membership in a common currency area.