高频数据下的金融波动率研究

发布时间：2018-01-04 13:18

本文关键词：高频数据下的金融波动率研究　出处：《西北师范大学》2013年硕士论文　论文类型：学位论文

【摘要】：在金融领域,不确定的风险即资产收益过程波动是市场逐利的根源,也是金融市场活跃的根本所在.资产收益过程在市场环境中受种种因素影响客观上形成不确定的波动,无论是出于逐利的目的或是资产保值的目的,人们总希望能够对即将到来的波动作出最为理想的预测,尤其在当今,利用先进的计算机技术和存储技术使得高频金融数据越来越容易得到,人们更希望使用对金融市场细节刻画更为细腻的高频金融数据对波动率作出更好的预测.因此,高频金融数据下对资产价格波动的研究越来越为人们所关注. 本文的主要工作和创新之处可以概括如下： 1.本文针对”日历效应”对波动率的影响这一问题,提出自加权波动的概念,通过自加权函数f来消除”日历效应”对波动率的影响,从而更准确地描述实际波动的特征.并从理论上证明了自加权积分波动率的极限性质,对波动率的应用具有指导意义. 2.本文提出用Bayes估计的方法对金融高频数据的波动率进行估计.通过利用以往信息,提出参数的先验分布,并对不同损失函数下的估计结果进行计算比较,从而得出一合理的估计结果. 3.在以往高频数据的研究中,观测时间点的选取往往都是等间距的选取.此种方法,首先并不能完全记录高频数据的变化趋势；其次,等间隔的时间点选取并没有考虑“日历效应”等对波动率的影响.本文针对以上问题,运用并改进了非等间隔选取观测时间点的方法.从而最大限度地保持原有数据的特性,使得最终估计出的波动率更加准确可靠. 4.金融高频数据中存在着微观结构噪声,并且微观结构误差随抽样频率的增加而增大.本文选用两个时间观测序列,将微观结构噪声带来的误差从波动率估计量中消除掉,从而使估计出的波动率更加真实可靠.
[Abstract]:In the field of finance, the uncertain risk, that is, the volatility of asset return process, is the root of market profit-seeking. Asset income process is affected by various factors in the market environment to form an objective uncertainty of fluctuations, whether out of the purpose of profit-seeking or the purpose of asset preservation. People always want to be able to make the most ideal prediction of the coming fluctuations, especially in today's, the use of advanced computer technology and storage technology to make high-frequency financial data more and more easily available. People prefer to use the finer high-frequency financial data to predict volatility. Therefore, the research on asset price volatility under the high-frequency financial data has attracted more and more attention. The main work and innovations of this paper can be summarized as follows: 1. Aiming at the influence of "calendar effect" on volatility, the concept of self-weighted volatility is proposed in this paper, and the influence of "calendar effect" on volatility is eliminated by self-weighting function f. The limit property of self-weighted integral volatility is proved theoretically, which is of guiding significance to the application of volatility. 2. In this paper, the Bayes estimation method is proposed to estimate the volatility of high-frequency financial data, and a priori distribution of the parameters is proposed by using the previous information. The estimation results under different loss functions are calculated and compared, and a reasonable estimation result is obtained. 3. In the previous research of high frequency data, the choice of observation time point is often equal distance selection. This method can not record the change trend of high frequency data completely. Secondly, the effect of "calendar effect" on volatility is not taken into account in the selection of equal-interval time points. This paper applies and improves the method of selecting observation time points at non-equal intervals, so as to maintain the characteristics of the original data to the maximum extent and make the final estimated volatility more accurate and reliable. 4. There is microstructure noise in financial high frequency data, and the microstructure error increases with the increase of sampling frequency. In this paper, two time series are selected. The error caused by microstructural noise is eliminated from the volatility estimator, so that the estimated volatility is more true and reliable.
【学位授予单位】：西北师范大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F830.9;F224

【参考文献】