In a seminal paper, Rose (2000) emphasises the strong effect of the currency union membership on trade, controlling for exchange rate volatility. This effect is described as puzzling by Rose himself. If capital markets are incomplete and reluctant to finance trade imbalances, attempts at trade integration might be constrained by the degree of bilateral business cycle correlation. A contrario, complete monetary integration within monetary unions allows a relaxation of the current account constraint, a phenomenon that explains Rose's estimate of extra trade within monetary unions, which swells to three times the normal volume of trade. This paper provides evidence that monetary and fiscal co-ordinations matter not only within monetary unions, but also within any monetary setting, from floats to pegs. Available data on government spending allow direct testing of hypotheses on policy co-ordination.