We test a simple model of exchange rate regime choice with data for 65 non-OECD countries covering the period 1980-94.We find that the variance of output at home and in potential target c ountries as well as the correlation between home and foreign real activity are powerful and robust predictors of exchange rate regime choice. Surprisingly, a more volatile foreign economy can be an argument in favor of a fixed exchange rate regime once similarities in the business cycle are taken into account. Comparable results hold for a variant of the model that focuses on nominal rather than real determinants. We also look at the impact of mistakes in exchange rate regime choice on actual (nominal) exchange rate volatility. Countries that deviate from the model\rquote s predicted regime by choosing fixed instead of floating exchange rates generally suffer higher exchange rate volatility than other countries having a fixed exchange rate regime. We also investigate the role of such mistakes in within\endash sample episodes of current-account crises.