基于G-H分布的谱风险度量分析及其Copula-SMR方法

发布时间：2018-04-26 11:52

本文选题：谱风险度量 + G-H分布　；参考：《北京化工大学》2012年硕士论文

【摘要】：在金融市场中，由于受到市场波动的影响，投资者和金融机构正面临着更加严峻的金融风险。如此一来，如何准确度量市场风险，来帮助投资者及金融机构的管理者规避风险，是经济学家们不断探索研究的主要工作。本文选用谱风险度量模型，，引入G-H分布，来作为对金融市场风险度量的主要方法。风险谱函数用来反应投资者对风险的厌恶程度，主要形式有三种，分别为指数型、幂型和双曲型风险谱函数。本文分别采用指数型和双曲型风险谱函数，选用G-H分布，结合谱风险度量方法，得到投资组合度量模型，并分别利用复化梯形法得到离散形式。对单支股票的风险度量分析，本文选取康美药业收益率序列作为样本数据，分别采用分位点估计法和矩估计法做G-H分布的参数估计，并分别与基于正态分布的谱风险度量模型进行对比，得到结论：选用G-H分布，并采用矩估计法估计参数，拟合效果最优。选用Kupiec失败率检验法，在一定置信水平下，利用样本内数据做谱风险度量分析，并与样本外数据进行比较，得到预测失败个数，得到结论。对于投资组合的风险度量分析，本文选取上海家化等四支证券收益率序列，分别采用指数型和双曲型风险谱函数，得到投资组合风险度量模型，得到结论：随着期望收益率的增加或绝对风险厌恶因子的增加，谱风险度量值也会相应增加，并且分配权重会向平均收益率高的证券转移。运用copula函数来度量不同证券之间相关性。选取万科A和海螺水泥两支证券，选取适当copula函数，建立投资组合模型。实证结果表明，由于考虑了证券收益的尾部相关性，可以得到更加贴近市场实际情况的风险度量结果。
[Abstract]:In the financial market, investors and financial institutions are facing more severe financial risks due to the influence of market fluctuations. Thus, how to measure the market risk accurately and help the investors and the managers of the financial institutions to avoid risk is the main work of the economists. This paper selects the spectrum risk measurement. The G-H distribution is used as the main method to measure the risk of financial market.
The risk spectrum function is used to reflect the degree of investor aversion to risk. There are three main forms, which are exponential type, power type and hyperbolic risk spectrum function. This paper uses exponential and hyperbolic risk spectrum functions, selects G-H distribution and combines spectral risk measurement method to obtain portfolio measurement model, and uses complex trapezoid method respectively. To the discrete form.
In the analysis of the risk measurement of single stock, this paper selects the recovery sequence of Kangmei pharmaceutical industry as the sample data, uses the point estimation method and the moment estimation method to estimate the parameters of the G-H distribution, and compares it with the spectral risk measurement model based on the normal distribution, and obtains the conclusion: the G-H distribution is selected and the moment estimation method is used to estimate the parameters. The fitting effect is optimal. The Kupiec failure rate test method is selected. Under certain confidence level, the spectrum risk measurement is analyzed by using the data in the sample, and compared with the data from the sample, the number of failure prediction is obtained and the conclusion is obtained.
For the risk measurement analysis of portfolio, this paper selects four stock return sequences, such as Shanghai domestication, and uses exponential and hyperbolic risk spectrum functions to get the portfolio risk measurement model. The conclusion is that as the expected return rate increases or the absolute risk aversion factor increases, the spectral risk measurement will also increase accordingly. And the distribution weights will be transferred to the securities with high average returns.
The copula function is used to measure the correlation between different securities. Two securities of Vanke A and conch cement are selected and appropriate Copula Functions are selected to establish an investment portfolio model. The empirical results show that the risk measurement results which are closer to the actual situation of the market can be obtained because of the tail correlation of stock returns.

【学位授予单位】：北京化工大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F832.51

【参考文献】