Nonparametric mean-lower partial moment model and enhanced index investment. (August 2022)

• Record Type: Journal Article
• Authors: Huang, Jinbo; Li, Yong; Yao, Haixiang
• Is Part Of: Computers & operations research. Volume 144(2022)
• Publication Date: 2022-08
• Subjects: C14 -- C61 -- G11 -- G23; Portfolio optimization -- Nonparametric kernel estimation -- Lower partial moment -- Enhanced index model; Operations research -- Periodicals; Electronic digital computers -- Periodicals; 004.05
• DOI: 10.1016/j.cor.2022.105814
• Languages: English
• ISSNs: 0305-0548

Abstract

Abstract: In this study, we propose a new smooth nonparametric kernel (NPK) method to estimate downside risk as measured by the lower partial moment (LPM) and build a NPK mean-LPM model. We also use the NPK LPM to construct a novel enhanced index model for controlling downside risk. We theoretically prove that the NPK LPM estimator is a convex function when the order of the LPM is equal to or greater than 1. The numerical results show that our NPK method outperforms the moment method in terms of estimation accuracy. Our simulated experiments and empirical tests show that the NPK LPM-based enhanced index model outperforms the benchmark models and indexes when using a wide variety of performance indicators and different rebalancing strategies. Highlights: We propose a new smooth nonparametric LPM to build a mean-LPM model. The nonparametric LPM is used to develop a novel enhanced index strategy. Theoretically the nonparametric LPM is a convex function of the portfolio position. The simulation results show that our method outperforms the moment method in terms of estimation accuracy. Empirical tests show our models outperform traditional indexation investment strategies.