Analyse factorielle dynamique: test du nombre de facteurs, estimation, et application a l'enquete de conjoncture dans l'industrie

Abstract

We suggest a two step procedure to estimate a dynamic factor model. We show that, in a stationary dynamic framework, static factor analysis leads to consistent estimators and allows to build an asymptotic test of the relevant number of factors. Once this number is set, the model can be estimated through a Kalman filter. We then apply this procedure to the French industrial business survey, in order to build a composite index.