The Portfolio Model Based On Regime Switch

In the investment practice of financial markets,it is crucial to use the reasonable inter-asset correlation to dynamically allocate the asset weights of the portfolio.This paper assumes that financial markets have different states which satisfy the homogeneous Markov chain.The GJR-GARCH model is used to estimate the dynamic volatility,and then the Regime Switching for Dynamic Conditional Correlations model is used to establish the covariance matrix estimation method.Finally,the estimated covariance matrix is applied to the traditional portfolio.Based on data from securities,bonds,and gold market,the empirical results show that the two states can reasonably reflect the high and low correlations between the markets,and the probability that the state lasts for a while is high.There is a significant difference in the correlation between the Shanghai Composite Index and the S&P 500,long-term bonds and SHFE gold in different market states.Finally,rebalancing through market states,this paper compares the portfolio performance based on the RS-DCC model with the portfolio under other multivariate GARCH models.The results show that the Maximum Sharpe Portfolio of RS-DCC model performs better under both revenue and risk.