论文部分内容阅读
当执行自上而下的资产配置策略时,地区资产选择是绕不开的步骤,对资产配置绩效影响巨大。使用我国32个地区板块指数2009-2014年间的日对数收益率数据,首先借助改进的指数分层结构算法对不同区域资产进行聚类,然后据此动态构建两组资产组合,并与随机资产组合结果进行对比分析。结果发现,指数分层结构算法较好刻画了区域资产之间的相关关系和拓扑结构特征,所得分类结果具有稳定性;无论是否控制行业因素,据此动态构建的跨区域资产组合都能快速降低风险,不仅能为机构投资者的积极资产管理提供有效指导,而且能更准确揭示市场的真实风险水平。
When implementing a top-down strategy of asset allocation, the choice of regional assets is an inaccessible step and has a huge impact on asset allocation performance. Using the daily logarithmic return data of 32 regions in China from 2009 to 2014, we first cluster the assets of different regions by means of improved exponential hierarchy algorithm, and then construct two groups of assets dynamically according to them, and compare them with random assets Combinatorial results for comparative analysis. The results show that the exponential hierarchy algorithm is better to depict the correlation between the regional assets and the topological features, and the classification results are stable. The dynamic cross-regional portfolio can be rapidly reduced with or without controlling the industry factors Risks not only provide effective guidance to the active asset management of institutional investors, but also reveal the real risk level of the market more accurately.