Heterogeneity in the Eurozone Monetary Policy Transmission Mechanism: A Structural BVAR Approach

Abstract

Since the inception of the European Monetary Union (EMU) in 1999, the issue of cross-country differences in monetary policy transmission has captured the interest of both economists and policymakers alike. We use a structural Bayesian vector autoregression (BVAR) model to investigate the diverging macroeconomic effects of monetary policy on the four largest Eurozone economies - Germany, France, Italy and Spain. We show that identification via heteroskedasticity, in conjunction with a large, multi-country BVAR model, identifies interpretable structural shocks without the imposition of strict restrictions on coefficients. Furthermore, estimation of a model on price and output data from 1999 to 2019 produces results contrary to the bulk of existing literature, finding a larger real output response in Spain and a smaller real output response in Germany. Upon estimation on a reduced sample from 1999 to 2014 that excludes the impacts of unconventional monetary policy, we find a reversal of the response strengths, corroborating the findings of prior studies. The change in cross-country rankings of response strength across different samples cannot be attributed to the evolution of structural characteristics - industry mix, export share, wage and employment rigidities - in these countries, suggesting the existence of new channels of policy transmission for UMP measures like quantitative easing. Principal components analysis of the error bands further reveal that the uncertainties of the long-run output responses for Germany and France, and Germany and Spain are negatively correlated. Overall, our findings have important implications for optimal policy-setting under asymmetric transmission and quantitative easing, and provide a new empirical framework for future studies of Eurozone macroeconomics.