Principal Components Estimation and Identication of the Factors

By Bai, Serena Ng. 2011

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

1. Jushan Bai
2. Serena Ng
3. Jianqing Fan
4. Kunpeng Li
5. Peng Wang
6. Lingzhou Xue
7. Jiawei Yao
8. Xu Han

Citations in the corpus

• Principal Components and Regularized Estimation of Factor Models (2017, background): “As as explained in Bai and Ng (2013), H NT will not, in general, be an identity matrix implying that the j -th factor estimate F j will not, in general, equal F 0 j even asymptotically. e The exception is when the true F 0 is such that F 0 ′ F 0 T = I r and Λ 0 ′ Λ 0 is diagonal.”
• Sufficient Forecasting Using Factor Models (2015, methodology): “Hence, with In addition, it is easy to verify that there is a constant C 3 such that By assumption, we also have Thus, Assumption A (c.ii) in Bai and Ng (2013) is satisfied with M 2 = C 3 M .”
• PROJECTED PRINCIPAL COMPONENT ANALYSIS IN FACTOR MODELS (2014, unknown): “It corresponds to the PC1 condition of Bai and Ng (2013), and is often used in the classical factor analysis for identification.”
• Identification and Estimation of Dynamic Factor Models (2012, methodology): “Define a dynamic factor model as in (4), (7), and (6).”
• TESTS FOR PARAMETER INSTABILITY IN DYNAMIC FACTOR MODELS (2013, background): “(2) Note that different behaviors of f̂tf̂ ′ t and ĝtĝ t result from the fact that both Ĝ1 and Ĝ2 do not have full column ranks under the alternative hypothesis.”
• What Drives Oil Prices? Emerging versus Developed Economies (2012, unknown): “To obtain unique identification, we follow the method proposed by Bai and Ng (2011).”
• Detecting big structural breaks in large factor models (2014, unknown): “Then, it can be shown that̂f1t = Op(1) and f̂2t = Op(δ−1N,T) for t = 1, . . . ,τ, so that1T ∑ τ t=1 f̂ 2 2t = Op(δ −2 N,T), and 1 T ∑ τ t=1 f̂1t f̂2t = Op(δ −2 N,T) (see Bai and Ng, 2011).”
• STATISTICAL ANALYSIS OF FACTOR MODELS OF HIGH DIMENSION (2012, background): “The results of [5] and [8] imply the following asymptotic representation for the principal components estimators.”