The Debate on Inflation Dynamics and the Swedish Experience

Sammanfattning: Inspired by the methodology of Galí and Gertler (1999) and Lindé (2005) and the seemingly fierce debate over the most appropriate econometric approach to study inflation dynamics, the new-Keynesian hybrid Phillips curve has been conducted with Swedish sample data and estimated with the general method of moment (GMM) and the full information maximum likelihood (FIML) estimators. GMM generated estimates that reject the idea of pure forward-looking behavior in price setting, but backwardness in price setting seems to be predominated. However, the generated results from GMM are somewhat conjectured in the sense that identification problem is apparent. Moreover the relationship between inflation and the so-called driving variable are not apparent. Due to this, a Granger-causality test between variables is employed. With identification in mind, the FIML yielded more significant estimates than GMM. FIML could not distinguish which price setting behavior is predominant to the same extent as the single equation estimator did. However, the estimated result seems to suggest that the majority of price setters have backward-looking behavior, indicating that backwardness in price setting has a relatively greater role. This is though the only aspect where the results in FIML are astray from the results in GMM, thus giving the impression that GMM is not obsolete. For instance, when it comes to the most relevant driving variable debate between marginal cost and output gap, the latter yielded estimates that has positive and correct sign. Indicating that output gap should also have an important role when studying Swedish inflation dynamics.