基于VaR模型的风险价值度量研究

发布时间：2018-03-17 11:44

本文选题：双曲分布　切入点：VaR　出处：《山东财经大学》2013年硕士论文　论文类型：学位论文

【摘要】：随着经济的全球化和金融的一体化，金融竞争与放松管制以及金融创新与技术取得了不断地进步，，金融市场和全球金融环境发生了重大的变化。与此同时，金融市场的系统风险和波动性也大大加剧，风险管理作为工商企业和金融机构的核心竞争力之一，其基础就是度量风险。VaR作为一种风险度量和管理的新工具，自诞生以来就得到广泛应用，目前在国外已成为度量市场风险的主流方法。传统的VaR计量方法存在两个问题，一是收益率分布的假设问题。传统的VaR计量是在正态分布的假设条件下进行的，然而正态分布并不能很好地描述收益率分布的尖峰厚尾的特点；二是对风险价值度量时往往笼统的运用收益率的全部历史数据进行分析，这就可能存在工作量大，度量结果不够精确的弊端。针对这两个问题，本文主要是从下面两个部分进行解决的。一是讨论了收益率分布的问题。本文分别运用描述统计量和拟合图对收益率是否符合正态分布做了初步研究，然后通过正态分布和双曲分布下VaR的度量结果平滑图比较，得出双曲分布能更好的描述收益率分布，从而否定了传统收益率正态分布的假设。二是对市场处于牛市和熊市情况下的风险价值进行度量。本文把收益率数据分成了上升趋势和下降趋势，首先用收益率的全部历史数据对牛市（熊市）情况下的风险价值进行度量，得到一组VaR值，然后再用上升趋势（下降趋势）的数据对牛市（熊市）情况下的风险价值进行度量，得到另一组VaR值，最后把这两组VaR值用平滑的曲线表示出来，比较用收益率全部的历史数据对牛市（熊市）情况下的风险价值度量和用上升趋势（下降趋势）的数据对牛市（熊市）情况下的风险价值度量的优劣程度，从而得出一种更有效的风险价值度量方法。
[Abstract]:With economic globalization and financial integration, financial competition and deregulation, financial innovation and technology have made continuous progress, and financial markets and the global financial environment have undergone major changes. As one of the core competitiveness of business enterprises and financial institutions, risk management is based on measuring risk. VaR is a new tool for risk measurement and management. It has been widely used since its birth, and has become the mainstream method to measure market risk in foreign countries. There are two problems in the traditional VaR metrology, one is the assumption of the return distribution, and the other is the assumption of the normal distribution. However, the normal distribution can not well describe the characteristics of the sharp and thick tail of the return distribution. Second, the risk value measurement is often generally used to analyze all the historical data of the yield, which may have a large workload. The drawback that the measurement result is not accurate. In view of these two problems, this paper mainly from the following two parts to carry on the solution. Firstly, the problem of yield distribution is discussed. In this paper, we use the descriptive statistics and the fitting graph to study whether the yield conforms to the normal distribution, and then compare the measured results of VaR under normal distribution and hyperbolic distribution. It is concluded that the hyperbolic distribution can better describe the return distribution, thus negating the hypothesis of the traditional normal distribution. The second is to measure the risk value of the market in bull market and bear market. This paper divides the yield data into upward trend and downward trend. First, we measure the value of risk in bull market (bear market) with all the historical data of yield, get a set of VaR values, and then measure the value of risk in bull market (bear market) with the data of upward trend (downward trend). You get another set of VaR values, and you end up with these two sets of VaR values represented by a smooth curve. To compare the degree of risk value measurement in bull market (bear market) and in bull market (bear market) with data of upward trend (downward trend) using all historical data of yield. Thus, a more effective method of risk value measurement is obtained.
【学位授予单位】：山东财经大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F830.91;F224

【参考文献】