将马尔科夫转换多分形模型引入均值-CVaR 框架, 构建资产组合优化的多分形模型, 并给出最优组合投资策略的求解步骤. 实证分析结果表明, 基于多分形模型的组合策略在描述性统计分析、风险调整收益分析、经济表现分析以及策略稳定性分析中的表现均优于基于 ~CCC-GARCH 模型的组合策略, 考虑多分形性确实有助于改善资产组合策略. 文章所构建的模型和实证结果为资产组合投资实践提供了有益的参考.

In this paper, we propose a multifractal model of portfolio optimization which incorporates Markov switching multifractal model into the mean-CVaR framework, and present the detailed solution procedures of optimal portfolio. The empirical results show that considering multifractality can improve portfolio investment strategy. According to the analysis of descriptive statistics, risk-adjusted rate of return, economic performance and stability of portfolio strategy, the investment strategy based on MSM model is superior to the one based on CCC-GARCH model. The proposed model and the associated empirical results are implicative for investment practice.