A0283
Title: A unified framework for estimation of high-dimensional conditional factor models
Authors: Qihui Chen - The Chinese University of Hong Kong, Shenzhen (China) [presenting]
Abstract: A general framework for estimating high-dimensional conditional latent factor models is presented via constrained nuclear norm regularization. Large sample properties of the estimators are established, and efficient algorithms are provided for their computation. To improve practical applicability, a cross-validation procedure is proposed for selecting the regularization parameter. The framework unifies the estimation of various conditional factor models, enabling the derivation of new asymptotic results while addressing the limitations of existing methods, which are often model-specific or restrictive. Empirical analyses of the cross-section of individual US stock returns suggest that imposing homogeneity improves the model's out-of-sample predictability, with the new method outperforming existing alternatives.
