VaR风险管理模型的理论与应用

【作者】 王良；

【导师】 李健元；

【作者基本信息】 东北财经大学 ， 金融学， 2003， 硕士

【摘要】 VaR风险管理模型作为一种符合未来风险管理发展方向的综合性风险度量方法，近年来得到了世界范围内主要商业银行、投资银行、基金管理公司及金融监管机构的普遍认可和支持，目前该模型已逐渐发展成为现代国际金融界风险度量与管理的主流方法。本文首先在简要介绍和分析VaR风险管理模型基本思想的基础上，基于VaR风险管理模型构建了统一的资产组合管理框架，探讨了VaR风险管理模型在资产组合管理中资产配置、风险管理和业绩评价的三大功能；其次，通过实证分析检验VaR风险管理模型在国内金融市场中的有效性；最后，在研究分析现代金融风险管理发展趋势和国内金融风险管理落后现状的基础上，对VaR风险管理模型在国内金融风险管理中的应用进行展望和分析。 通过研究分析，本文主要得出如下结论：(1)传统的Markowitz均值——方差模型仅仅是在资产组合收益率正态分布假设条件下基于VaR风险管理模型进行资产组合选择的特例，与均值——方差模型中的方差风险度量方法相比，VaR风险管理模型能够更全面、更贴切地衡量资产组合的风险，且基于此模型能够更有效地进行资产配置决策；(2)VaR风险管理模型能够满足更高层次风险管理者对风险信息的需求，有助于整体风险管理效率的提高；(3)基于VaR风险管理模型的RAROC绩效评价能够反映资产组合管理人的真实业绩，从而为金融机构风险限额的分配和激励约束机制的制定提供统一的标准；(4)国内证券市场资产组合收益率服从正态分布的假设明显不成立，实证检验表明基于资产组合收益率正态分布假设条件下的方差——协方差模型对国内资产组合风险的预测存在较大的偏差，由于文中证明在收益率正态分布假设条件下基于方差——协方差模型进行资产组合选择的结果等价于Markowitz的均值——方差模型，因此，均值——方差模型对国内资产组合风险的预测同样会存在着较大的偏差，而半参数VaR风险管理模型则能够取得较好的预测衡量效果；(5)VaR风险管理模型符合未来金融风险管理的发展趋势，基于VaR风险管理模型建立内容提要风险限额内控体系、风险信息披露体系和业绩评价体系，并进行金融监管，将有助于国内金融机构内部风险管理方法和外部监管技术跟上国际金融风险管理的发展潮流。更多还原

【Abstract】 As an integrated risk-measuring approach, the VaR Risk Management Model is in accord with the new trend of risk management development. In recent years, VaR Risk Management Model has been generally accepted by the main commercial banks, investment banks, fund management companies and the institutions of the financial supervision. Today, this model has become the most popular risk-measuring approach in the world. Starting from the fundmental principle of the VaR Risk Management Model, this paper firstly constructed a portfolio management framework which is based on this model, and analyzed three functions of this model in the framework: asset allocation, risk management and performance valuation. Then, this paper empirically tested the validation and predictive accuracy of different VaR Risk Management Model in the domestic financial market.Finally,with the analysis of modem financial risk management development trend and the current domestic financial risk management situation, this paper made a prospect for the application of this model in the construction of domestic financial risk management system.Through the analysis, the main conclusions are as follows:(l)The traditional Mean-Variance Model is the special example of the portfolio selection based on the VaR Risk Management Model for the case that the returns of the portfolio are assumed to be normally distributed; Compared with the Mean-Variance Model ,the VaR Risk Management Model is more comprehensive and accurate in the measurement of the portfolio risk, so based on the VaR Model, the investors can allocate the asset more effectively. (2)The VaR Risk Management Model can provide the timely and comprehensive risk information for the top risk manager, so it is very helpful to the improvement of total risk management efficiency.(3)Based on the VaR Model, the RAROC performance valuation approach can reflect the real performance of the portfolio manager and provide the coherent standard for the allocation of risk limitation and the construction of the incentive compatibilityconstraint mechanism in the financial instiutions. (4)The normal distribution assumption of the domestic financial asset return is evidently incorrect. There are too many unacceptable biases in the Variance-Covariance Model’s prediction of future risk. Because under the assumption of normal distribution, the Variance-Covariance Model has the same selection of portfolio as the Mean-Variance Model, so the prediction of the Mean-Variance Model has the same biases, however the semi-parameter VaR Model could give a better prediction. (5) The VaR Risk Management Model is complying with the new trend of the financial risk management development. Based on the VaR Model, constructing the risk-limitition internal control system, risk information disclosure system, performance valuation system and strengthening the financial supervision will help the technology of domestic financial risk management to catch up with the world’s level.更多还原