Will US in ation awake from the dead? The role of slack and non-linearities in the Phillips curve

摘要

The response of US in ation to the high levels of spare capacity during the Great Recession of 2007-09 was rather muted. At the same time, some have argued that the short-term unemployment gap has a more prominent role than the standard unemployment gap in determining in ation, and either the closing of this gap or non-linearities in the Phillips curve could lead to a sudden pick-up in in ation. In this context, our main aim is to provide guidance to policymakers as regards the reliability of the Phillips curve to forecast in ation. Our main findings from Phillips curves estimated since the early-1990s suggest that the consideration of a time-variation in the Phillips curve slope is more relevant than just focusing on-finding the "correct" slack measure. Although the Phillips curve may be relatively at over the full sample (1992Q1-2015Q1), time varying estimates with rolling windows and with the Kalman filter suggest that the slope does vary over time and that it has increased slightly since 2013. These non-linear specifications outperform the benchmark linear model in an out-of-sample exercise. The main policy implication is that decision-makers should not exclude the possibility that in ation might rise suddenly given its non-linear behaviour, and more strongly than a linear model would dictate. (C) 2017 The Society for Policy Modeling. Published by Elsevier Inc. All rights reserved.