Dynamic factor models: does the specification matter?

Thumbnail Image
Publication date
Defense date
Journal Title
Journal ISSN
Volume Title
Google Scholar
Research Projects
Organizational Units
Journal Issue
Dynamic Factor Models (DFMs), which assume the existence of a small number of unobserved underlying factors capturing the comovements in large systems of variables, are very popular among empirical macroeconomists to reduce dimension and to extract factors with an economic interpretation. Factors can be extracted using either non-parametric Principal Components (PC) or parametric Kalman filter and smoothing (KFS) procedures, with the former being computationally simpler and robust against misspecification and the latter being efficient if the specification is correct and coping in a natural way with missing and mixed-frequency data, time-varying parameters, non-linearities and non-stationarity among many other stylized facts often observed in real systems of economic variables. This paper analyses the empirical consequences on factor estimation and forecasting of using alternative extraction procedures and estimators of the DFM parameters under various sources of potential misspecification. In particular, we consider factor extraction when assuming different number of factors and different factor dynamics. The factors are extracted from a popular data base of US macroeconomic variables that has been widely analyzed in the literature without consensus about the most appropriate model speciffication. We show that this lack of consensus is ony marginally cruzial when it comes to factor extraction but it matters when the objective is forecasting.
Em Algorithm, Kalman Filter, Principal Components, State-Space Model
Bibliographic citation