Modelling euro area GDP growth with dynamic factor models

Using structural models to estimate and forecast quarterly GDP conditional on a large
number of economic indicators raises parsimony problems, such as the need to define many
identification restrictions and also to deal with very sparse asymptotic distributions of the
coefficients. Factor models have the role to fill this gap in macroeconomic modelling. The
basic assumption is that the correlation between the observed variables can be sufficiently
well explained by a few common orthogonal factors. The estimated factors can then be
used to estimate any variable of interest. In the empirical section of this paper, using the
Area Wide Model database, we obtain near-term-forecasts (NTF) of euro area GDP using
the dynamic factor model. The specification of the model is such that we can decompose
the deviations of the NTF’s from the long run mean, which can be interpreted in a similar
manner as the output gaps.