【作者】 史宇峰；

【导师】 张世英；

【作者基本信息】 天津大学 ， 技术经济及管理， 2008， 博士

【摘要】 金融资产收益相关性及持续性研究是金融工程研究领域的中心,准确度量资产收益之间的相关性及持续性是探索资本市场运行机理和实务操作的关键。为此,将统计物理学和多元统计学中的有关理论及金融计量模型与金融领域的现实问题相结合取得了如下成果:1.以统计物理学中的随机相关矩阵特征值概率密度为理论依据,构建了收益相关矩阵选取方法,从而解决了收益相关矩阵优选问题。实证检验显示,按由该方法所选取的收益相关矩阵配置资产可得到优质的有效前沿。2.以多元统计学中的Wishart概率密度为理论依据,推导出相关系数的修正模型,该模型剔除了样本容量效应,使相关性度量更为准确可靠。实证结果显示,该模型具有较高的资产配置效率。3.基于协同持续思想通过GARCH模型技术构建了动态投资组合模型,该模型可捕捉到金融资产收益的局部波动,达到控制风险扩散的目的。实证显示,按该模型配置资产可控制组合收益率在较小范围内波动,也可得较高夏普比。4.提出风险溢出发生期及风险溢出强度两个概念,并构建了相应数学模型。同时,基于该模型进行了实证研究,结果显示,沪市影响深市的程度较深市影响沪市的程度大,且它们之间的风险溢出在大约3分钟之内完成,这一研究结果与实际吻合。5.将统计物理学中的公因子提取法运用于资本资产定价方面,提出了CAPM修正模型。实证结果显示该模型较原模型具有较高的定价能力。上述理论成果依次在论文五个主体章节中展开讨论,这些研究成果均遵循从理论依据到方法构建、模型提出,再到实证检验的研究思路,从而确保了研究工作的科学性和实用性。更多还原

【Abstract】 It is most important step to study on the correlation and persistence of the financial capital return,especially,is most crucial work measuring accuratelly the correlation and persistence to explore the mechanism and operation of the capital market. To this end, by applying the theory of statistical physics and multiple-statistics and the model for financial measures to the practical problems of financia field, the achievements are as follows:1. The method to discerning the best correlation matrix is obtainded by the random matrix eigenvalue probability density function in statistical physics solving the matrix optimization. Empirical tests show that the efficient frontier may be bettler by the method .2. Based on Wishart probability distribution of multiple statistics, the revised model of the correlation coefficient is derived so that the sample pool doubt disappear and so the correlation coefficient becomes an accurate measure tool.3. After considering the Co-persistence thought, the dynamic portfolio model is built based on GARCH model,and the model can capture the local fluctuations of the financial assets,which is helpful to control the risk proliferation. The smaller volatility of portfolio return and the higher Sharp ratio of the allocation are showed by simulating the financial data.4. The spillover period concept and the spillover strength concept of the risk are raised,and the corresponding mathematical models are producted. At the same time,the empirical research reflect that Shanghai financial market impacts more deeply on the Shenzhen market,in contrary to,the Shenzhen market more weakly impacts Shanghai.And the 3 minute spillover period is reflected.The results are consistent to the fact.5.The common factor extraction method of Statistical physics is used in the capital asset pricing in order that the revised CAPM are invented. The model is more capable than the original model in capital asset pricing. These theoretical results were in five main sections of the paper, and, these studies were processed by following from the basis theory to the methods and the models, and ,last fact test ,so ,the metioned outcomes are scientific and practicality.更多还原