基于网络搜索数据的突发事件对石油价格波动的影响研究

发布时间：2018-04-29 12:37

本文选题：突发事件 + 油价波动　；参考：《武汉大学》2017年硕士论文

【摘要】：随着经济的飞速发展,石油在国民经济中的地位也越来越突出,近年来,各类突发事件频发,石油价格呈现出极其剧烈的波动。通过对以往突发事件的影响进行分析,得出规律性的一致变化特征,可以为新时代下的油价走势预测提供经验,提高我国石油突发事件的风险应对能力。随着互联网技术的普及,人们的许多传统观念被改变和颠覆,目前我们的社会已经进入信息化技术高速发展的大数据时代,互联网逐渐成为人们搜索信息、了解各式信息的全新工具。搜索引擎记录了用户的搜索行为,网络搜索数据反映了用户对事件的关注度,网络搜索数据凭借低成本、及时性为各行各业的预测提供了很多帮助,已经被广泛地运用于社会经济行为的预测之中,并取得了良好的效果,其可行性得到证实。基于此,本文将网络搜索数据纳入到分析框架之中,首先分析了目前中国石油市场的供给、需求、进口现状和面临的风险问题,在此基础上探讨了突发事件对油价波动的影响机制,为后续的分析奠定基础;然后,基于结构性断点检验对国际石油价格走势和波动特征进行分析;再次,基于网络搜索数据对传统事件分析法进行改进,探讨历史上的四次典型突发事件对事件窗内油价超额收益率是否有显著影响,并分析其和石油价格波动的关系;最后,建立自回归分布滞后、广义自回归条件异方差等模型,利用特定突发事件的网络关注度指标对石油价格波动进行解释,对其影响力大小进行有效衡量与监测。研究发现:重大突发事件对国际石油市场有着显著的短期影响,对事件窗内的累计超额收益率影响显著;市场关注能够加剧突发事件对石油市场的冲击作用,四个突发事件的网络搜索关注度均对油价产生显著影响,卡特里娜飓风和利比亚战争的市场关注度正向作用于油价,而美国次贷危机和墨西哥湾漏油的影响为负向,美国次贷危机有着最大的影响程度和持续时间;利比亚战争期间的油价波动具有显著的GARCH效应,利比亚战争对石油市场影响的半衰期为1～2天,且该影响程度呈边际递减趋势。
[Abstract]:With the rapid development of economy, the status of petroleum in the national economy is becoming more and more prominent. Through the analysis of the influence of the past unexpected events, the consistent change characteristics of the regularity can be obtained, which can provide experience for the prediction of oil price trend in the new era and improve the ability to deal with the risks of oil emergencies in China. With the popularity of Internet technology, many traditional concepts of people have been changed and subverted. At present, our society has entered the era of big data with the rapid development of information technology, and the Internet has gradually become a search for information. New tools for understanding all kinds of information. The search engine records the user's search behavior, the network search data reflects the user's attention to the event, the network search data provides a lot of help for the forecast of various industries by virtue of its low cost and timeliness. It has been widely used in the prediction of social economic behavior, and has achieved good results, and its feasibility has been proved. Based on this, this paper brings the network search data into the analysis framework. Firstly, it analyzes the supply, demand, import status and risk of China's oil market. On this basis, the influence mechanism of unexpected events on oil price fluctuation is discussed, which lays the foundation for further analysis. Then, based on structural breakpoint test, the trend and volatility characteristics of international oil price are analyzed. Thirdly, This paper improves the traditional event analysis method based on the network search data, discusses whether the four typical emergencies in history have a significant impact on the oil price excess return rate in the event window, and analyzes its relationship with the oil price fluctuation. An autoregressive distribution lag model and generalized autoregressive conditional heteroscedasticity model are established to explain the fluctuation of oil price by using the network concern index of a particular unexpected event and to measure and monitor its influence effectively. It is found that major emergencies have a significant short-term impact on the international oil market, and a significant impact on the cumulative excess return rate in the event window; market concern can aggravate the impact of the unexpected events on the oil market. The Internet search focus on the four emergencies has a significant impact on oil prices. Hurricane Katrina and the Libyan war market attention is positively affecting oil prices, while the US subprime mortgage crisis and the Gulf of Mexico oil spill have a negative impact. The oil price fluctuation during the Libyan War has significant GARCH effect, and the half-life of the Libyan war's impact on the oil market is 1 ~ 2 days, and the influence degree shows a marginal decreasing trend.
【学位授予单位】：武汉大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F416.22;F764.1

【相似文献】