基于VaR-GARCH模型和分位数回归的国际石油市场对人民币汇率市场的风险溢出效应研究

发布时间：2017-12-31 20:34

本文关键词：基于VaR-GARCH模型和分位数回归的国际石油市场对人民币汇率市场的风险溢出效应研究　出处：《浙江财经大学》2014年硕士论文　论文类型：学位论文

【摘要】：近年以来，在金融全球化的助推下国际政治经济舞台越来越复杂，同时考虑到国际石油市场与国际石油价格的猛烈变化，国际石油市场的风险管理已经成为了全球关注的焦点。如果说石油作为一种最原始的推动力，是现代社会前进和运行的最基本保证，那么汇率是另外一个受到国际社会各界关心的焦点。和国际石油价格一样，汇率也是各国经济发展的重要因素。作为一种基础性的能源大宗商品，国际石油的价格变化对世界各国的经济发展有着极其重要的影响，在一定程度上也决定着汇率的变化。改革开放以来，我国的经济发展水平明显提高，人民的生活也不断改善。而这些社会现象出现的背后是我国对石油需求量不断提升的残酷现实。最近几年，我国的石油需求量几乎一直以每年两位数的速度不断增长。我国已经成为世界上石油需求量最大的国家之一。进口大量石油不仅要花费大量的外汇储备，而且高比例的进口依存度也对未来我国的石油安全造成了隐患。然而在这样不利的情况下，我国还没有得到国际石油的定价权，中东地区的石油是我国最重要的进口石油来源，而中东地区不断的战乱经常导致石油的剧烈波动。我国的经济发展和金融稳定会受到国际石油市场上油价波动的一定影响，最显然的是在用外汇进口石油的过程中，国际石油价格波动产生的风险必然引起人民币汇率市场产生对应的风险溢出效应。因此，分析国际石油价格波动对人民币汇率市场的冲击有着重要的研究价值。由于近年来国际石油市场的准金融属性不断提高，使用金融的方法度量国际石油价格风险已经成为学术界的主题之一。而金融风险往往是由于金融资产价格的波动所产生的，因此，如果可以较为准确地测量出资产价格的波动率，那么人们就可以度量出资产或资产组合将会产生的潜在风险。使用GARCH簇模型可以比较好的测量出资产价格的波动率，为准确测量金融风险奠定了基础。 VaR作为一种全新的风险度量方法，具有方法直观、结果量化、结论易懂等优点，相比传统的风险度量方法，例如情景模拟法、压力测试法、灵敏度方法等更受到风险管理者的喜爱，也受到许多学者的青睐。随着VaR模型的不断改善，其在金融风险管理中的应用越来越广泛，同时也符合金融机构和学术界对风险进行集中化、数量化管理的要求。考虑到GARCH簇模型和VaR与CoVaR方法的优点，将以上方法从度量金融市场风险的角度引入到度量国际石油市场风险的角度不失为一种崭新的研究思路。本文首先介绍了和论文有关的国内外研究现状和发展趋势，其次阐述了石油市场风险管理的相关理论，，然后讲述了GARCH簇模型和VaR与CoVaR模型的基本原理，最后是论文的核心部分——实证分析。在实证部分，论文首先对对Brent石油市场和WTI石油市场的对数收益率两个时间序列的统计特点(偏度、峰度、JB值等)进行分析，发现两个市场的对数收益率序列具有尖峰厚尾性、自相关性和ARCH效应，因此可以对这两个市场使用GARCH簇模型做实证分析。于是，论文假设在偏t分布的分布条件下，使用GARCH簇模型来描述Brent石油市场和WTI石油市场的石油价格波动情况，并对各个GARCH模型进行参数估计，得到对应的波动率。其次，在使用GARCH簇模型得到Brent石油市场和WTI石油市场的市场价格波动率的同时，利用非参数的方法计算出两个对数收益率的残差序列的风险值VaR，并使用半参数的方法计算出两个石油市场的VaR值，对其进行回测检验和模型预测未来风险。最后，鉴于两个石油市场的实际收益率数据具有尖峰厚尾、异方差性和波动集聚性等特点，为了提高参数估计的准确性，注重度量参数在分布上尾和下尾的影响，本文使用分位数回归的方法来计算国际石油市场对人民币汇率市场产生的条件在险值CoVaR及对应的风险溢出大小。
[Abstract]:In recent years, in boosting the financial globalization under the stage of international politics and economy are becoming more and more complex, taking into account the drastic changes in the international oil market and international oil prices, the risk management of the international oil market has become the focus of global attention. If the oil is one of the most original driving force, is the most basic guarantee of modern social progress and the operation, then the exchange rate is also a focus of international concern from all circles of society. And the international oil price, the exchange rate is also an important factor in the economic development of countries. As a kind of basic energy commodities, has an extremely important impact on economic development and changes in the international price of oil in the world, in a certain degree also decides the change of exchange rate.
Since the reform and opening up, China's economic development level is obviously improved, people's lives are constantly improving. Behind these social phenomenon is the cruel reality of China's demand for oil increasing. In recent years, the demand for oil in China has almost two per year double-digit increase. China has become one of the largest countries in the world oil demand. Import a lot of oil not only to spend a large amount of foreign exchange reserves, and a high proportion of import dependence on petroleum security of our country. However, the hidden dangers in such adverse circumstances, our country has not obtained the international oil pricing right, Middle East oil is China's most important sources of oil imports, while volatility constant war often lead to oil in the Middle East. The development of China's economic and financial stability will be affected by the international oil market prices Some effects of fluctuations, the most obvious is in the process of foreign exchange for oil imports, risk of international oil price fluctuations will inevitably cause the risk spillover effect corresponding to the RMB exchange rate market. Therefore, the analysis of international oil price fluctuations on the impact of RMB exchange rate market has important research value.
Due to the international oil market in recent years the quasi financial attributes increasing method using financial measure of international oil price risk has become one of the topics in academic circles. The financial risk is often due to the fluctuation of financial assets generated, therefore, if we can accurately measure the asset price volatility, so people can to measure the potential risk will be asset or portfolio generated. Using the GARCH cluster model can better measure the asset price volatility, which laid the foundation for the accurate measurement of financial risks.
VaR as a new risk measure method, with intuitive method, quantitative results, conclusion to understand the advantages, compared with the traditional risk measurement methods, such as the scene simulation method, pressure test, sensitivity method is more risk managers alike, but also by many scholars of all ages. With the continuous improvement of the VaR model. And the financial risk management in China is widely used, but also in line with the centralized risk of financial institutions and the academic circles, the number of management requirements.
Considering the advantages of GARCH cluster model and VaR and CoVaR method, it is a new research idea to introduce the above methods from the perspective of measuring financial market risk to measuring the risk of international oil market.
This paper first introduces the research status and development trend related to papers, and then expounds the relevant theories of risk management in oil market. Then it describes the basic principles of GARCH cluster model and VaR and CoVaR models, and finally, the core part of the paper, empirical analysis.
In the empirical part, firstly the statistical characteristics of log return on Brent WTI oil market and oil market rate of two time series (skewness, kurtosis, JB value etc.) are analyzed, found the log return two market rate series has a fat tail, autocorrelation and ARCH effects, so we can do the empirical analysis using the GARCH cluster model of the two market. Therefore, the assumptions on the distribution conditions of partial t distribution, using GARCH cluster model to describe the oil price fluctuations of oil market and oil market Brent WTI, and the GARCH model to estimate the parameters, get the corresponding volatility. Secondly, Brent and oil market WTI oil market volatility of market price at the same time using the GARCH cluster model, using non parametric methods to calculate the risk of two log return residuals value VaR method is used to calculate the parameters of half Two oil market value of VaR on the back test and predict future risk. Finally, in view of the actual yield data of two oil market has peak thick tail, heteroscedasticity and volatility clustering characteristics, in order to improve the accuracy of parameter estimation, focusing on the metric parameters and lower tails in the distribution, we use quantile regression method to calculate the international oil market of RMB exchange rate market conditions and the corresponding value of CoVaR in the size of Risk Spillover risks.

【学位授予单位】：浙江财经大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F224;F416.22;F832.6

【参考文献】