基于CAPM-SV-Copula模型的投资组合研究

发布时间：2018-03-19 11:37

本文选题：市场因素　切入点：SV　出处：《南京理工大学》2014年硕士论文　论文类型：学位论文

【摘要】：Markowitz均值-方差投资组合选择理论自创立以来,在经济金融研究领域中备受推崇,取得了一系列极为深刻的理论结果和广泛的实际应用效果。周知,Markowitz均值-方差模型的构建依赖于诸如资产的预期收益分布已知且满足独立同分布、投资者单期效用最大化等假设条件。但随着经济生活环境的不断变化和发展,这些假设条件或过于理想或很难满足。因之,寻找更为合理、有效、符合实际金融市场状况的广义投资组合(可能有别于Markowitz收益线性投资组合)模型及分析方法,显得自然而然且自有其理论推广和现实应用的双重含义与价值。 GARCH-Copula模型乃至SV-Copula模型的构建,为这一广义投资组合想法的实现提供了思路和手段,为解决资产收益非独立同分布的组合投资管理提供了一个初步的可行的方案。但是,可能是因为这两个模型的数理分析上的难度使然,两模型都忽略了一个基本事实：市场因素对广义投资组合的影响。有鉴于此,本文充分考虑了市场因素对广义组合投资的影响,深入研究并构建了基于市场因素影响下的CAPM-SV-Copula模型,以此测度金融风险,并为投资者的资产组合选择管理提供一个合理、有效的分析工具。首先,本文考虑了市场因素对波动率和收益率的影响,将CAPM模型和SV模型相结合,构建了广义CAPM-SV模型,并对工商银行股票收益率序列进行了实证分析。研究发现本文提出的广义CAPM-SV模型在数据拟合效果和风险预测能力方面比原来SV模型的效果要好。其次,将广义CAPM-SV模型与Copula方法结合,构建了CAPM-SV-Copula模型,并对中国工商银行和中国石油构成的投资组合进行实证分析。结果表明考虑了市场因素而构建的模型在风险刻画能力方面,特别是描述尾部风险能力方面,优于传统的SV-Copula模型的刻画效果。
[Abstract]:Since Markowitz's mean-variance portfolio selection theory was founded, it has been highly respected in the field of economic and financial research, and has obtained a series of profound theoretical results and extensive practical application results. It is well known that the construction of Markowitz's mean-variance model depends on assumptions such as that the expected return distribution of assets is known and satisfies the independent same distribution, and that investors maximize single-period utility. However, with the continuous changes and development of the economic living environment, These assumptions are either too ideal or difficult to meet. As a result, look for more reasonable, efficient, broadly based portfolio models and analytical methods that are likely to be different from Markowitz return linear portfolios, in line with actual financial market conditions. It is natural and has its dual meaning and value of theoretical popularization and practical application. The construction of the GARCH-Copula model and even the SV-Copula model provides the idea and means for the realization of this generalized portfolio idea, and provides a preliminary feasible scheme for solving the portfolio investment management, which is not independent and distributed in terms of asset returns. Perhaps because of the difficulty in mathematical analysis of the two models, both models ignore a basic fact: the influence of market factors on the broad portfolio. In this paper, the influence of market factors on generalized portfolio investment is fully considered, and the CAPM-SV-Copula model based on market factors is studied and constructed in order to measure financial risk and provide a reasonable choice management for investors' portfolio. Effective analytical tools. Firstly, considering the influence of market factors on volatility and yield, the generalized CAPM-SV model is constructed by combining CAPM model with SV model. The empirical analysis of ICBC stock return series shows that the generalized CAPM-SV model proposed in this paper is more effective than the SV model in data fitting and risk prediction. Secondly, the generalized CAPM-SV model is combined with the Copula method to construct the CAPM-SV-Copula model. The empirical analysis of the portfolio of ICBC and PetroChina shows that the model, which takes market factors into account, has the ability to describe the risk, especially the tail risk. It is better than the traditional SV-Copula model.
【学位授予单位】：南京理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F830.59;F224

【参考文献】