Research on Short-Term Returns of Chinese Asset Portfolios Based on Markov Regimes

Abstract

This study constructs a dynamic weighting strategy based on the Markov Regime-Switching (MS-VAR) model to forecast five-day returns and optimize investment portfolios in the Chinese multi-asset market. The research selects seven representative assets: the CSI 300 Index, Shanghai Composite Index, STAR Market 50 Index, CSI Bank Index, Kweichow Moutai, Contemporary Amperex Technology Co. Limited (CATL), and Ping An Insurance. The sample period spans 3,741 trading days from January 4, 2010, to December 31, 2024. By introducing CVaR constraints and a maximum drawdown stop-loss rule, the MS-VAR strategy significantly outperformed traditional strategies in out-of-sample tests in terms of risk-adjusted returns, particularly excelling during extreme events like the 2016 circuit breaker, the 2018 trade war, the 2020 pandemic, and the 2022 lockdown periods. The study finds distinct regime-switching characteristics in the Chinese multi-asset market, demonstrating that dynamic allocation based on state probabilities can effectively control tail risks and enhance investment performance.