Abstract
Conventional Phillips-curve models that are used to estimate the output gap detect a substantial decline in potential output due to the present crisis. Using a multivariate state space model, we show that this result does not hold if the long-run role of excess liquidity (that we estimate endogenously) for inflation is taken into account.
Acknowledgments
The author is indebted to Oliver Holtemöller, Sebastian Giesen, Henry Dannenberg and Peter Haug for valuable comments and discussions.
Notes
1However, this does not necessarily imply that the production capacity is fully utilized. Contrarily to the original definition, production potential is not the production capacity in these models but rather a trend component that captures the production at the hypothetical average level of capacity utilization. Thus, the gap can be positive in these models in times of a boom.
2Stationarity was tested using augmented Dickey–Fuller, Phillips–Perron and KPSS tests.
3Although we do not model the income elasticity of money demand explicitly, the above definition of (business cycle neutral) velocity does not imply a restriction to an income elasticity of money demand equal to 1, because all persistent components of velocity are removed independent of their origin. This data-driven approach allows a parsimonious specification that is essential to derive meaningful results using a Kalman filter.