Model selection in factor-augmented regressions with estimated factors
针对因子增强回归，提出两种有限样本下一致的模型选择方法：留d出交叉验证和基于bootstrap的预测误差平方近似，并通过模拟和美股溢价实证验证其性能。
This paper proposes two consistent model selection procedures for factor-augmented regressions (FAR) in finite samples. We first demonstrate that the usual cross-validation is inconsistent, but that a generalization, leave-d-out cross-validation, is consistent. The second proposed criterion is a generalization of the bootstrap approximation of the squared error of prediction to FARs. The paper provides the validity results and documents their finite sample performance through simulations. An illustrative empirical application that analyzes the relationship between the equity premium and factors extracted from a large panel of U.S. macroeconomic data is conducted.