A0797
Title: Nonlinear IPCA with instruments selection
Authors: Qi Zhang - Southwestern University of Finance and Economics (China) [presenting]
Qiuhua Xu - Southwestern University of Finance and Economics (China)
Abstract: The estimation and variable selection are studied in semi-parametric conditional factor models with a single-index structure. The model employs the sieve method and singular value decomposition (SVD) for estimation and conducts variable selection on the characteristics variables. The large-N asymptotic properties of the factor estimators and the large-NT asymptotic properties of the factor loadings and pricing error estimators are established. The consistency and Oracle properties of the adaptive LASSO are also demonstrated in variable selection. Additionally, a weighted bootstrap procedure is developed for testing whether the pricing errors are zero. Finally, an empirical study is conducted on the U.S. stock market, presenting the predictive results and the selected characteristics variables.
