TY - JOUR
T1 - Bayesian factor-adjusted sparse regression
AU - Fan, Jianqing
AU - Jiang, Bai
AU - Sun, Qiang
N1 - Funding Information:
Fan and Jiang's research was supported by NSF grant DMS-1662139 and NIH grant 2R01-GM072611-14.
Publisher Copyright:
© 2021
PY - 2022/9
Y1 - 2022/9
N2 - Many sparse regression methods rely on an assumption that the covariates are weakly correlated, which hardly holds in many economic and financial datasets. To relax this assumption, we model the strongly correlated covariates by a factor structure: strong correlations among covariates are modeled by common factors, while the remaining variations of covariates are modeled as idiosyncratic components. We then propose a factor-adjusted sparse regression model and develop a semi-Bayesian estimation method for it. Posterior contraction rate and model selection consistency are established by a non-asymptotic analysis. Experimental studies show that the proposed method outperforms its Lasso analogue, manifests insensitivity to overestimates of the number of common factors, pays a negligible price when covariates are uncorrelated, scales up well with increasing sample size, dimensionality and sparsity, and converges fast to the posterior distribution. An application to the U.S. bond risk premia lends further support to the proposed model and method.
AB - Many sparse regression methods rely on an assumption that the covariates are weakly correlated, which hardly holds in many economic and financial datasets. To relax this assumption, we model the strongly correlated covariates by a factor structure: strong correlations among covariates are modeled by common factors, while the remaining variations of covariates are modeled as idiosyncratic components. We then propose a factor-adjusted sparse regression model and develop a semi-Bayesian estimation method for it. Posterior contraction rate and model selection consistency are established by a non-asymptotic analysis. Experimental studies show that the proposed method outperforms its Lasso analogue, manifests insensitivity to overestimates of the number of common factors, pays a negligible price when covariates are uncorrelated, scales up well with increasing sample size, dimensionality and sparsity, and converges fast to the posterior distribution. An application to the U.S. bond risk premia lends further support to the proposed model and method.
KW - Bayesian sparse regression
KW - Factor model
KW - Model selection
KW - Posterior contraction
UR - http://www.scopus.com/inward/record.url?scp=85118535356&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85118535356&partnerID=8YFLogxK
U2 - 10.1016/j.jeconom.2020.06.012
DO - 10.1016/j.jeconom.2020.06.012
M3 - Article
C2 - 35754940
AN - SCOPUS:85118535356
SN - 0304-4076
VL - 230
SP - 3
EP - 19
JO - Journal of Econometrics
JF - Journal of Econometrics
IS - 1
ER -