基于二阶嵌套模拟法估计VaR的研究

发布时间：2018-02-04 23:26

本文关键词： 风险价值蒙特卡罗模拟二阶嵌套模拟条件期望方差　出处：《北京化工大学》2013年硕士论文　论文类型：学位论文

【摘要】：风险是金融市场不可避免的问题，随着金融市场进一步全球一体化，投资者们迫切需要更精确、更便捷的风险信息。VaR（风险价值）自被提出以来，一直作为金融机构与投资者最常用的金融工具。风险价值能全面量化复杂投资组合的风险，已成为金融界热点研究问题。本文基于蒙特卡罗模拟法计算风险价值的基础上，结合近几年提出的二阶蒙特卡罗模拟，将其结合中国金融实际环境，提出两种新的计算风险价值的方法。论文中首先概述选题的背景和意义、国内外文献综述、研究的内容、方法和创新之处，接着介绍了VaR模型和利用蒙特卡罗模拟法计算VaR的方法，，随后介绍了金融资产收益率的统计学特征以及传统蒙特卡罗模拟法计算VaR的缺陷与不足，并在此基础上介绍二阶蒙特卡罗模拟，即嵌套模拟。我们分别提出用基于嵌套模拟的自适应分配算法和基于嵌套模拟估计条件期望方差的方法来计算VaR，在理论上整理和论证了样本抽取和误差分析以及方差分析的渐近理论等，结合中国国内实际的股票数据进行实证研究，并分别对其模型进行有效性检验。本文的结论是：利用嵌套模拟来估计风险，能有效解决传统蒙特卡罗模拟法的弊端，两种新的、高效的计算VaR的方法在中国金融市场实证研究中得到了很好的印证。本文针对国内金融市场首次利用2阶蒙特卡罗模拟来研究金融风险，并提出了能充分利用计算机性能的自适应分配算法和根据条件期望的方差来估计VaR的方法，这两种基于嵌套模拟的计算方法对中国股票市场的风险估计都有很不错的结果。
[Abstract]:Risk is an inevitable problem in financial market. With the further globalization of financial market, investors urgently need more accurate and convenient risk information. As the most commonly used financial instrument for financial institutions and investors, the value of risk can comprehensively quantify the risk of complex portfolio, and has become a hot research issue in the financial field. In this paper, based on the Monte Carlo simulation method to calculate the value of risk, combined with the second order Monte Carlo simulation proposed in recent years, combined with the actual financial environment in China. This paper first summarizes the background and significance of the topic, literature review at home and abroad, research content, methods and innovations. Then it introduces the VaR model and the method of using Monte Carlo simulation to calculate VaR, and then introduces the statistical characteristics of the return on financial assets and the defects and shortcomings of the traditional Monte Carlo simulation method for calculating VaR. On this basis, we introduce second-order Monte Carlo simulation, that is, nested simulation. We propose an adaptive allocation algorithm based on nested simulation and a method based on nested simulation to estimate conditional expected variance to calculate VaR. In theory, the paper collates and demonstrates the asymptotic theory of sample sampling, error analysis and variance analysis, and carries out empirical research based on the actual stock data in China. The conclusion of this paper is that using nested simulation to estimate risk can effectively solve the disadvantages of traditional Monte Carlo simulation. The efficient method of calculating VaR is well confirmed in the empirical study of Chinese financial market. In this paper, the second order Monte Carlo simulation is used to study financial risk for the first time in the domestic financial market. An adaptive allocation algorithm which can make full use of computer performance and a method to estimate VaR based on conditional expected variance are proposed. The two methods based on nested simulation have good results for the risk estimation of Chinese stock market.
【学位授予单位】：北京化工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F224;F830.9

【参考文献】