A two-step estimator for large approximate dynamic factor models based on Kalman filtering

Catherine Doz (1) , Lucrezia Reichlin (2)

1 PSE - Paris School of Economics
2 London Business School

This paper shows consistency of a two step estimation of the factors in a dynamic approximate factor model when the panel of time series is large ( large). In the first step, the parameters of the model are estimated from an OLS on principal components. In the second step, the factors are estimated via the Kalman smoother. The analysis develops the theory for the estimator considered in Reichlin, and Sala (2004) and Giannone et al. (2008) and for the many empirical papers using this framework for nowcasting.

Mots clés: C51 C32; C33; Factor models; Kalman filter; Principal components; Large cross-sections

Domaines: Econométrie de la finance [q-fin.ST]

Catherine Doz, Lucrezia Reichlin. A two-step estimator for large approximate dynamic factor models based on Kalman filtering. Econometrics, 2011, 164 (1), pp.188. ⟨10.1016/j.jeconom.2011.02.012⟩. ⟨hal-00844811⟩