The empirical “gravity” equation is extremely successful in explaining bilateral trade. This paper shows how a multi-country model of specialization and costly trade (i.e. a microfounded gravity model) can be applied to explain empirical exchange rate puzzles. One such puzzle is the fact that nominal exchange rates are enormously volatile, but that this volatility does not appear to affect inflation. The gravity model is very successful in explaining this puzzle. In a sample of 25 OECD countries in the post- Bretton Woods period, the gravity prediction of inflation substantially outperforms the purchasing power parity prediction. The gravity prediction matches the volatility of actual inflation, and tracks its path closely. The superior performance of the gravity prediction is explained primarily by the fact that it takes account of the interaction of specialization with home bias. The stability of inflation in very open economies is explained in addition by the fact that the size of bilateral trade is negatively correlated with bilateral exchange rate volatility.
