Abstract
We estimate the Euro Area output gap by means of a large-dimensional non-stationary dynamic factor model, exploiting the information stemming from a large panel of macroeconomic series. Estimation is performed by means of a three-steps technique which explicitly accounts for the COVID shock, combined with a Principal Component Analysis on the estimated factors to decompose aggregate output in a trend and a cyclical component, namely potential output and the output gap. We find that COVID has a significant impact both on levels and on the variability of the latent shocks, thereby reinforcing the need to explicitly model such features. The estimated gap is in line with the institutional estimates, with the main differences arising in the aftermath of the Great Recession. The proposed procedure complements standard model-based approaches with a fully data-driven estimate.