Official and four alternative regime classification schemes based on observed exchange rate behaviour are used to examine the relationship with inflation and growth in developing countries. For an identical sample of observations from 73 countries for 1984-2001, only the scheme based on parallel rates suggests a significant effect (negative) of floating on growth. Floats that claim to be pegs, or have high exchange rate volatility, are the ones with lower growth. Hard pegs offer inflation benefits. Floating is not consistently associated with higher inflation than soft pegs, and any apparent association is a possible by-product of the design of the classification algorithms.