Gravity models have been shown to be fairly effective in modelling bilateral trading patterns, explaining more than 50 per cent of the variation in trade. This paper examines bilateral trade patterns using a data set provided by Rose and van Wincoop (2001). Rose (2000) has suggested that forming a currency union has a dramatic effect on the volume of intra-union trade. A number of econometric issues are identified with respect to this claim. There is some evidence that Rose's (2000) empirical results are not entirely robust to the sample of countries used, and to the estimation method. In particular, some of the regressors may be endogenous, which casts doubt on the magnitude of the parameter estimates.