Large-Dimensional Dynamic Factor Models in Real-Time: A Survey

Abstract

In this paper I review the literature on Large-Dimensional Dynamic Factor Models for real-time applications. I first present the Dynamic Factor model, the implications of using large-dimensional databases, and the challenges of real-time applications. Then, I discuss how the literature has solved these problems, and I present numerous empirical applications that show the usefulness of these models in both constructing business cycle indicators, and predicting economic activity. Finally, I present two recent extensions of the Dynamic Factor model, one in a Bayesian and one in a non-stationary setting.

Keywords: Large Dimensional Dynamic Factor models, Real-time forecasting, business cycle indicators

JEL Classification: C32, C53, E32