A spectral EM algorithm for dynamic factor models

A spectral EM algorithm for dynamic factor models

Series: Working Papers. 1619.

Author: Gabriele Fiorentini, Alessandro Galesi and Enrique Sentana.

Published in: Journal of Econometrics, Volume 205, Issue 1, July 2018, Pages 249-279Opens in new window

Full document

PDF
A spectral EM algorithm for dynamic factor models (818 KB)

Abstract

We make two complementary contributions to efficiently estimate dynamic factor models: a frequency domain EM algorithm and a swift iterated indirect inference procedure for ARMA models with no asymptotic efficiency loss for any finite number of iterations. Although our procedures can estimate such models with many series without good initial values, near the optimum we recommend switching to a gradient method that analytically computes spectral scores using the EM principle. We successfully employ our methods to construct an index that captures the common movements of US sectoral employment growth rates, which we compare to the indices obtained by semiparametric methods.

Previous Spanish public finances thr... Next Fiscal multipliers across t...