Published online by Cambridge University Press: 23 June 2017
We employ a non-recursive identification scheme to identify the effects of a monetary policy shock in a Structural Vector Autoregressive (SVAR) model for the US post-WWII quarterly data. The identification of the shock is achieved via heteroskedasticity, and different on-impact macroeconomic responses are allowed for (but not imposed) in each volatility regime. We show that the impulse responses obtained with the suggested non-recursive identification scheme are quite similar to those conditional on a recursive VAR estimated with pre-1984 data. In contrast, recursive vs. non-recursive identification schemes return different short-run responses of output and investment during the Great Moderation. Robustness checks dealing with a different definition of investment, an alternative break-point, and federal funds futures rates as an indicator of the monetary policy stance are documented and discussed.
We thank William A. Barnett (Editor), an anonymous Associate Editor, and two anonymous referees for insightful comments, and Giovanni Caggiano, Riccardo Lucchetti, Michel Normandin, Giovanni Pellegrino, and Jouko Vilmunen for valuable discussions. We also thank Christopher Crowe for kindly providing us with the new shock series proposed in the Barakchian and Crowe (2013, Journal of Monetary Economics) paper. Participants to presentations held to the Fourth International Conference in memory of Carlo Giannini 2014 (Pavia), the CFE conference 2014 (Pisa), the Society for Computational Economics 2014 (Oslo), the International Association for Applied Econometrics Conference 2015 (Thessaloniki), the Bank of Finland and the University of Padova provided us with useful feedback. Emanuele Bacchiocchi and Luca Fanelli gratefully acknowledge partial financial support from the Italian MIUR Grant PRIN-2010/2011, prot. 2010RHAHPL 003; Luca Fanelli also acknowledges RFO grants from the University of Bologna.