Factor models are widely used in macroeconomic forecasting. With large datasets, factor models are particularly useful due to their intrinsic dimension reduction. In this chapter, we consider the forecasting problem using factor models, with special consideration to large datasets. In factor model estimation, we focus on principal component methods, and show how the estimated factors can be used to assist forecasting. Machine learning methods are discussed to encompass the high-dimensional features of large factor models. We consider policy evaluation as a nowcasting problem and show how factor analysis can be used to perform counter-factual outcome prediction in complicated models with observational data. The usage of all these techniques is illustrated by empirical examples.
CITATION STYLE
Cao, J., Gu, C., & Wang, Y. (2020). Principal Component and Static Factor Analysis. In Advanced Studies in Theoretical and Applied Econometrics (Vol. 52, pp. 229–266). Springer. https://doi.org/10.1007/978-3-030-31150-6_8
Mendeley helps you to discover research relevant for your work.