基于VAR的风险管理方法比较及在证券市场中的实证检验

发布时间：2018-04-16 07:46

本文选题：VaR方法 + 正态分布　；参考：《辽宁大学》2013年硕士论文

【摘要】：证券市场风险的度量，即对证券市场风险进行测量，是证券市场风险管理的基础。需要构建合适的模型使用恰当的方法，而这一方面的研究也是当前金融研究领域的一个热门话题。证券市场风险度量的模型和估计方法是多种多样的，其中使用的最为广泛的是VaR（Value-at-Risk）方法。VaR是在正常的市场环境下，在某一特定的时间区间和置信度水平，对预期的最大损失进行测量的一种方法。VaR方法适用于复杂的投资组合，说明该投资组合的杠杆效应和分散效果。首先它可以给证券持有者提供风险量化指标，指导内部决策的制定；其次在进行投资决策时，VaR还可以对收益和预期风险进行权衡。VaR方法提供了一种关于市场风险的综合性度量，是建立在可靠的科学基础之上的。在理论方面，本文对VaR方法做了较详尽的介绍。先对VaR的计算方法在不同层面作了详细的介绍，其中包括在正态分布、t分布以及GED分布下的VaR计算；后又详细介绍了各种方差预测模型，，本文在前人研究的基础上，综合了所有模型进行介绍；之后的第三章，对各种计算方法和方差预测模型分别进行了较全面的理论比较，其中还简单介绍了另外两种常用的风险度量方法，并与VaR方法作了简单的比较。在实证检验方面，我们以上证综指为样本数据，把不同分布下的其中八种最有效的模型应用到实证中，通过数据输出结果的比较，在24中模型中通过分析参数估计，进行VaR计算结果检验。实证的结果显示：与正态分布与t分布相比，GED分布能够更好的刻画上证综指的波动性，且在GED分布下，各种方差预测模型中PARCH模型得出的检验值结果最好。
[Abstract]:The measurement of securities market risk, that is, the measurement of securities market risk, is the basis of securities market risk management.It is necessary to construct the appropriate model and use the appropriate method, which is also a hot topic in the field of financial research.There are a variety of models and estimation methods for risk measurement in securities market. The most widely used method is VaRN Value-at-Risk.VaR is in a normal market environment, in a specific time interval and confidence level.The method of measuring the expected maximum loss. VaR method is suitable for complex portfolio, which shows the leverage effect and dispersion effect of the portfolio.First, it can provide securities holders with quantitative risk indicators to guide the making of internal decisions; secondly, VaR can also weigh the return and expected risks to provide a comprehensive measure of market risk.Is based on sound science.In theory, the VaR method is introduced in detail.Firstly, the calculation methods of VaR are introduced in detail at different levels, including the calculation of VaR under normal distribution t distribution and GED distribution, and then various variance prediction models are introduced in detail.After the introduction of all the models, in the third chapter, the author makes a comprehensive theoretical comparison of various calculation methods and variance prediction models, and briefly introduces the other two commonly used risk measurement methods.A simple comparison is made with the VaR method.In the empirical test, we take the Shanghai Composite Index as the sample data, and apply eight of the most effective models under different distribution to the empirical results. Through the comparison of the data output results, we analyze the parameter estimation in the 24 model.The results of VaR calculation were tested.The empirical results show that: compared with normal distribution and t-distribution, PARCH distribution can better describe the volatility of Shanghai Composite Index, and under the GED distribution, PARCH model has the best test results.
【学位授予单位】：辽宁大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F830.91;F224

【参考文献】