高频金融数据波动率计算方法的研究

发布时间：2018-04-02 03:18

本文选题：高频金融数据　切入点：波动率　出处：《长春工业大学》2013年硕士论文

【摘要】：高频金融数据是指日内的金融时间序列,是以小时、分钟或秒为抽样频率的数据。高频金融数据样本容量大,采集周期短,比低频数据包含了更丰富的日内收益波动信息,能够更好的反映金融市场特征。随着计算机技术与通讯技术的飞速发展高频金融数据的获取、存储、检索和处理已经十分便利。为了更深入地研究金融市场的微观结构,对高频金融数据波动率的研究已经成为国内外金融领域的研究热点。各国的经济学家们已经对高频金融数据波动率进行了比较深入的研究。波动率是衍生工具定价、投资组合以及金融风险管理的关键变量,构建准确的金融波动估计量具有重要的理论和实践意义。Andersen与Bollerslev在1998年提出一种全新的波动率估计的计算方法——“已实现”波动(RV),该方法无模型,计算简便,在一定条件下是波动率的一致估计量,为这一领域的研究提出了新思路。为了得到更有效的估计方法,国内外学者又对“已实现”波动(RV)进行了改进和扩展,给出了多种计算方法。本文首先介绍了研究背景、选题的意义以及主要的工作,然后介绍了金融高频数据研究的现状,并指出了目前研究中存在的问题。在此基础上研究了金融高频数据波动率计算方法,针对“已实现”波动(RV)和“已实现”双幂次变差(RBV)这两个高频金融波动率的计算方法进行了介绍,并做了比较,指出“已实现”双幂次变差(RBV)在定义形式上比“已实现”波动(RV)所包含的内容更加广泛,除了具有稳健性,还证明了在两种条件下,“已实现”双幂次变差(RBV)比“已实现”波动(RV)更有效。并以定理的形式指出“已实现”双幂次变差是比“已实现”波动更有效的波动率估计量。同时又对波动率的计算方法进行了改进和扩展,通过对深证成指和上证综指2011年12月14日到2012年12月14日以1分钟为采样周期的收盘价高频金融数据进行了实证分析,实证分析表明,改进和扩展的波动率计算方法是更有效的。
[Abstract]:High frequency financial data financial time series is in the days, hours, minutes or seconds for the sampling frequency data. High frequency financial data samples of large capacity, short cycle of data acquisition, more days volatility contains information than low frequency data, which can reflect the characteristics of financial market better. With the rapid development of high frequency acquisition the financial data of computer technology and communication technology for the storage, retrieval and processing have been very convenient. In order to further study the microscopic structure of the financial market, the research of high frequency financial data volatility has become a hot research topic at home and abroad in the financial sector.
Economists, countries have conducted deep research on high frequency financial data. The volatility of volatility is the pricing of derivatives, portfolio and key variables of financial risk management, construction of the accurate estimation of the amount of financial volatility has important theoretical and practical significance of.Andersen and Bollerslev proposed a new calculation method in volatility estimation 1998 -- "realized volatility (RV), the method of model free, simple calculation, under certain conditions is consistent with the estimation of volatility, put forward new ideas for the research in this field. In order to get a more effective estimation method, scholars at home and abroad on realized volatility (RV) the improvement and expansion, given a variety of calculation methods.
This paper first introduces the research background, the significance of the topic and the main work, then introduces the research status of financial high frequency data, and points out the existing problems in the current study. On the basis of the research on the volatility of high frequency data calculation method for realized volatility (RV) and realized the double power variation (RBV) method to calculate the two high frequency financial volatility are introduced and compared, pointed out that the realized bipower variation (RBV) in the form of definition than the realized volatility (RV) contains more widely, in addition to the robustness, also proved in the two condition, the realized bipower (RBV) than the realized volatility (RV) is more effective. And in the form of the theorem that the realized bipower variation is greater than the realized volatility of more effective volatility estimator. At the same time wave Calculating method of dynamic rate was improved and extended by the Shenzhen and Shanghai from December 14, 2011 to December 14, 2012 to 1 minutes for the closing price of high frequency financial data sampling period by empirical analysis, empirical analysis shows that the calculation method of the improvement and extension of the volatility is more effective.

【学位授予单位】：长春工业大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F224;F830.9

【参考文献】