基于Copula方法在开放式基金投资组合风险管理中的应用研究

发布时间：2018-01-08 03:28

本文关键词：基于Copula方法在开放式基金投资组合风险管理中的应用研究　出处：《湖南大学》2012年硕士论文　论文类型：学位论文

【摘要】：针对传统的相关系数矩阵并不能很好的描述金融资产间的非线性相依结构，本文将时变Copula函数与GARCH模型相结合，建立Copula-GARCH模型研究开放式基金投资组合的风险和相依结构。该模型可以有效规避由于传统线性相依结构假定所带来的模型设定误差，同时还考虑了相依性的动态性变化，以便于更好的刻画资产间相依结构的时变性特征。本文首先对开放式基金的概念以及内涵进行阐述，同时比较已有的度量市场收益风险的方法。其次应用自相关检验对开放式基金净值收益时间序列进行检验，以判别开放式基金收益时间序列是否存在高阶序列相关性。并应用ARCH-LM检验对该时间序列进行检验，判别是否具有高阶ARCH效应，并以此为依据，选择适当的GARCH类模型对开放式基金收益的边缘分布进行描述。为了表示所假定的分布能对开放式基金收益时间序列的分布特征进行较好的刻画，应用K-S检验以及A-D检验对开放式基金收益序列边缘分布模型的拟合优度进行检验。再次，运用不同的时变Copula函数对开放式基金资产间的相依性关系进行描述。最后，将蒙特卡罗模拟技术与时变Copula函数相结合，对开放式基金组合风险进行VaR测度，以反映开放式基金投资组合风险变化情况与其资产在不同风险情况下可能损失的概率。我们选取了有代表性的三支开放式基金作为研究对象，，发现时变Clayton Copula-AR(p)-GARCH模型所估计的VaR值更为准确，并且能够在各个置信水平下较全面地覆盖最大损失风险，从而达到控制风险的目的。本文的创新主要体现在：所构建的模型能有效地避免金融时间序列所普遍存在的尖峰厚尾以及有偏性特征所造成的模型设定误差；本文通过拟合优度检验确定最优Copula类型，以确保实证结果的稳健性以及准确性。
[Abstract]:Because the traditional correlation coefficient matrix can not well describe the nonlinear dependent structure between financial assets, this paper combines the time-varying Copula function with the GARCH model. Copula-GARCH model is established to study the risk and dependent structure of open-end fund portfolio. This model can effectively avoid the model setting error caused by the traditional linear dependent structure assumption. At the same time, the dynamic variation of dependency is considered in order to better describe the time-varying characteristics of the dependent structure of assets. This paper first describes the concept and connotation of open-end funds, and compares the existing methods to measure the market return risk. Secondly, we use autocorrelation test to test the time series of net income of open-end funds. In order to determine whether the open-end fund income time series has higher order sequence correlation, and use ARCH-LM test to test the time series, judge whether there is high order ARCH effect. And based on it. Select appropriate GARCH model to describe the marginal distribution of open-end fund income. In order to express the assumed distribution of open-end fund income time series distribution characteristics can be well described. K-S test and A-D test are used to test the goodness of fit of the edge distribution model of open-end fund income series. Third. Different time-varying Copula functions are used to describe the relationship between open-end fund assets. Finally, Monte Carlo simulation technology and time-varying Copula function are combined. The risk of open-end fund portfolio is measured by VaR. In order to reflect the risk change of open-end fund portfolio and the probability of possible loss of its assets under different risk conditions, we select three representative open-end funds as the research object. It is found that the estimated VaR value of the time-varying Clayton Copula-AR(p)-GARCH model is more accurate and can cover the maximum loss risk more comprehensively at various confidence levels. In order to achieve the purpose of risk control. The innovations of this paper are as follows: the model constructed can effectively avoid the model setting error caused by the ubiquitous peak and thick tail of financial time series and the bias characteristic; In order to ensure the robustness and accuracy of empirical results, the optimal Copula type is determined by goodness of fit test in this paper.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F830.59

【参考文献】