• calibration
• dsge
• high order perturbation methods
• high order perturbation methods
• independent component analysis
• macroeconomics
• model confidence set
• non guassianity
• svar identification

I discuss a new method for the calibration and validation of DSGE models. The
calibration is performed following a procedure which exploit different instruments.
The core tool is a the model confidence set (MCS), which allows to find the set of
the "best" configurations of parameters exploiting a minimum distance index (MDI).
The MCS at the same time allows also to validate the calibration. The MDI measure
the distance between the matrix of contemporaneous relations of the SVAR applied
to the real data, and applied to the data obtain by the simulations of the model. The
SVAR is identified through independent component analysis theory, which is based
on the non-gaussianity of the shocks of the models. The thesis proposes a detailed
description of the model used, the chosen solution method, the literature review of
other method of calibration, the methodology applied and the result obtained.