A number of theoretical models predict that the slope of the Phillips curve increases with trade openness, but cross-country studies provide little evidence for such a correlation. We highlight two reasons for this …nding. Firstly, the strength of the relationship may depend on the extent of exchange rate adjustment, which is a potential determinant of output and in ‡ation dynamics in open economies, but previous studies have not made a distinction between …xed and ‡oating exchange rate regimes.Secondly, existing estimates of the Phillips curve slope are based on data from the 1950s through the 1980s, and are therefore likely a¤ected by price and wage controls, in ‡ationary oil price hikes and the role played by …scal policy in driving output and in ‡ation (the underlying theory requires that monetary shocks dominate). We calculate new measures of the Phillips curve slope using data from 1981-98, a period during which these factors were arguably less important. Regressions based on the new measures indicate that the Phillips curve slope increases with trade openness amongst countries maintaining ‡exible and semi- ‡exible exchange rate regimes, but is unrelated to openness amongst countries maintaining …xed exchange rate regimes.