Effects of monetary policy regime changes in the Euro Economy

There prevails a view that the monetary policy in the Euro area has changed, especially after the European Monetary System (EMS) came into operation in 1979. The goal of this paper is to quantify (1) the extent to which policy regime has changed and (2) the counterfactual effects of such regimes changes, using a multiple-equation dynamic model. We first apply the constant-parameter version of the model to the two subsamples of quarterly data: 1970-1979 and 1980-2002. We find that both the systematic part of monetary policy and impulse responses to policy shocks are different across both subsamples, pointing to potential regime switching in policy. But because sampling errors of the estimates for the two subsamples are large (due to the short sample size), it is impossible to tell how different the two regimes really are. To test if policy regimes are different and to quantify the impact of such a regime change, it is essential to restrict the time variations in both the monetary policy equation and the rest of the system. We follow the Sims and Zha (2004) method to identify monetary policy regimes endogenously; We separate the policy behavior from the private-sector behavior to test whether regime change occurs in the private sector as well. The study conducted here is important because constant-parameter vector autoregressive (VAR) models have been used as a benchmark by Smets and Wouters (2003) to select the best DSGE model by adding shocks. If the economic structure underlying observed statistical correlations among data series has fundamentally changed over time, the constant-parameter VAR model is deemed inadequate. We will use the Bayesian approach to compare a litany of models with different degrees of time variations and select the one that has the best fit to data. Such a model will provide a better benchmark for macroeconomists to test their DSGE models for the Euro economy.