Footnote:
Nach Informationen von SSRN wurde die ursprüngliche Fassung des Dokuments October 11, 2021 erstellt
Description:
Conventional measurements of risk premiums are biased if the estimation models are potentially misspecified and unstable. Say, factor interactions is one of the crucial omitted specifications that standard models cannot involve. Motivated by this argument, we propose an interpretable factorization-based method to estimate the risk premium of factors in a linear asset pricing model (we call it Factorization Asset Pricing Model, FAPM), which is able to account for all interactions between factors using factorized parameters. We emphasize the critical importance of the factor interactions in measuring risk premiums. We show that our factorization approach can be identified as the best-performing method among current methodologies (including trees and neural networks, among other nonlinear models), even in a parsimonious linear framework. We also highlight that few factors input can predict well, while numerous factors set may generate negative effects due to adverse factor interactions. Remarkably, weak factors in standard models may play important roles in FAPM because their interactions with other factors can be significant