Inflation Dynamics, Marginal Cost, and the Output Gap: Evidence from Three Countries

Recent studies have argued that the New Keynesian Phillips curve (Calvo pricing model) is empirically valid, provided that real marginal cost rather than detrended output is used as the variable driving inflation. One interpretation of this result is that real marginal cost is not closely related to the output gap, and so models for monetary policy need to include labormarket rigidities. An alternative interpretation is that marginal cost and the output gap are closely related, but that the latter needs to be measured in a manner consistent with dynamic general equilibrium models. To date, there has been little econometric investigation of this alternative interpretation. This paper provides estimates of the New Keynesian Phillips curve for the U.S., the U.K., and Australia using theory-consistent estimates of the output gap. Using this theory to measure the output gap leads to a considerable improvement in the empirical performance of output-gapbased Phillips curves.