Working Paper: CEPR ID: DP3285
Authors: Lucrezia Reichlin
Abstract: This Paper reviews recent econometric work on factor models in large cross-sections of time series. In this literature, traditional factor analysis is adapted to develop parsimonious estimation methods for high dimension time series models. The review covers problems of consistency and rates ? as the dimension of the cross-section and the time dimension become large ? identification and forecasting. We also review empirical applications on measuring and interpreting business cycles.
Keywords: business cycles; factor analysis; panel data
JEL Codes: C22; C23; E32; E37
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| common shocks (u) (Y50) | individual variables (x) (C29) |
| number of time series (n) (C32) | identification of common shocks (E32) |
| macroeconomic shocks (q) (E39) | observable through individual variables (C29) |
| common shocks (u) (Y50) | economic behavior (D22) |