Decomposing the Asset Pricing Anomalies: Evidence from China

Abstract

This paper introduces a functional principal component analysis (FPCA) to decompose China’s A-share portfolio returns on time-series and cross-section simultaneously. The results show that the first empirical functional principal component (EFPC) stands for the market factor and the others for an anomaly. The second and third ones reveal the cross-sectional linear and convex patterns, and the joint of them dominates the asset pricing anomalies. Furthermore, the EFPCs illustrate much more information than the portfolio-based approach, and we can use them to explain the debates about some anomalies.