基于SV类模型的国际油轮运价指数波动风险研究

发布时间：2018-03-09 12:04

本文选题：国际油轮运价指数　切入点：随机波动模型(SV)　出处：《上海交通大学》2012年硕士论文　论文类型：学位论文

【摘要】：近年来,作为体现国际油轮运输市场运价走势的运价指数波动越来越剧烈,而航运界对运价指数波动风险的研究主要集中在干散货市场,传统的油轮运价指数波动风险度量方法已经无法满足新竞争形势下的要求,因此从理论上寻找更优的模型来度量国际油轮运价指数波动风险,更加充分的认识其波动性,从而为油轮运输企业进行风险控制提供更好理论依据显得十分重要。随机波动(SV)模型在刻画金融序列波动特征方面优于传统的广义自回归条件异方差(GARCH)模型,因此本文尝试引入SV模型来刻画国际油轮运价指数波动及其特征,结合广义帕累托分布GPD分布的极值理论,建立起度量国际油轮运价指数波动风险的动态VaR模型,主要内容如下: 首先主要介绍了国际油轮运价指数的概况、构成、计算和主要特征,指出油轮运价指数波动的尖峰厚尾性,波动集聚性,长记忆性和持续性,以及对正、负冲击产生不同回应的杠杆效应,为后文做模型的对比分析奠定基础。其次从理论上分别介绍了三类SV和GARCH模型,即标准模型,厚尾模型和杠杆模型,重点阐述了SV模型参数估计的MCMC方法和模型判别的DIC准则,在此基础上结合基于GPD的极值理论和VaR方法建立起动态风险度量模型,并给出模型适用性的失败率验证方法。最后选取一定数量BDTI和BCTI指数进行实证研究,分别构建SV模型和GARCH模型,并对两模型进行对比分析,指出SV类模型在刻画油轮运价指数波动特征方面的优势,得出L-SV模型为最优模型,而T分布条件下,GARCH-T模型最优的结论,后采用L-SV模型进行最终的动态VaR值计算,失败率验证结果表明L-SV模型能够很好的度量出油轮运输市场的波动风险,特别适用于原油运输市场,并且国际原油运输市场的波动风险要大于国际成品油运输市场。
[Abstract]:In recent years, as a reflection of the trend of the international tanker transportation market, the fluctuation of the freight rate index is becoming more and more intense, while the research on the risk of the fluctuation of the freight rate index in the shipping industry is mainly focused on the dry bulk cargo market. The traditional risk measurement method of oil tanker price index fluctuation has been unable to meet the requirements of the new competitive situation, so we can find a better model to measure the fluctuation risk of international tanker price index in theory, and fully understand its volatility. Therefore, it is very important to provide better theoretical basis for oil tanker transportation enterprises to carry out risk control. The stochastic volatility model is superior to the generalized autoregressive conditional heteroscedasticity (GARCH) model in describing the volatility characteristics of financial series. Therefore, this paper attempts to introduce SV model to describe the volatility and its characteristics of international tanker freight index. Combined with the extreme value theory of the generalized Pareto distribution GPD distribution, a dynamic VaR model is established to measure the volatility risk of the international tanker price index. The main contents are as follows:. First of all, the general situation, composition, calculation and main characteristics of the international tanker freight index are introduced. It is pointed out that the peak, thick tail, agglomeration, long memory and persistence of the fluctuation of the oil tanker price index are positive. The leverage effect of different responses is produced by negative shock, which lays a foundation for the comparative analysis of the later models. Secondly, three kinds of SV and GARCH models are introduced theoretically, that is, standard model, thick tail model and lever model. The MCMC method of SV model parameter estimation and the DIC criterion of model discrimination are discussed in detail. On this basis, a dynamic risk measurement model is established by combining the extreme value theory based on GPD and the VaR method, and the failure rate verification method for the applicability of the model is given. Finally, a certain number of BDTI and BCTI indices are selected for empirical research, SV model and GARCH model are constructed, and the two models are compared and analyzed, and the advantages of SV model in characterizing the fluctuation characteristics of tanker price index are pointed out. The conclusion that L-SV model is the optimal model and the GARCH-T model is optimal under the condition of T distribution is obtained. Then the L-SV model is used to calculate the final dynamic VaR value. The result of failure rate verification shows that the L-SV model can measure the fluctuation risk of tanker transportation market very well. It is especially suitable for the crude oil transportation market, and the fluctuation risk of the international crude oil transportation market is greater than that of the international refined oil transportation market.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F551;F224

【引证文献】