A1788
Title: Data-driven identification and estimation of DSGE models by non-Gaussianity
Authors: Damiano Di Francesco - Sant\'Anna School of Advanced Studies (Italy) [presenting]
Alessio Moneta - Scuola Superiore Sant'Anna (Italy)
Mario Martinoli - Sant Anna School of Advanced Studies (Italy)
Raffaello Seri - University of Insubria (Italy)
Abstract: A new procedure is proposed to estimate the parameters of a dynamic stochastic general equilibrium (DSGE) model from observed macroeconomic time series. The approach combines impulse response function matching with indirect inference and statistical identification by non-Gaussianity, using vector autoregressive (VAR) models as auxiliary models. A key element of the approach is a minimum distance index whose argument is a pair of impulse response function matrices. This index serves two objectives. First, it allows for the identification of structural impulse response functions from the data generated by the DSGE model, exploiting the non-Gaussianity of the observed data but allowing Gaussianity of the model-simulated data. Second, it enters as an objective function in the indirect inference procedure. The proposed procedure is illustrated by applying it to a simple new Keynesian DSGE model.
