| 题名 | 基于高频数据的金融市场风险测度研究; Measurement of Financial Market Risk Based on High Frequency Data |
| 作者 | 苗晓宇 |
| 答辩日期 | 2011 ; 2011 |
| 导师 | 曾五一 |
| 关键词 | （超）高频数据 (Ultra) High Frequency Data 金融市场风险 Financial Market Risk 测度方法 Measure Method 已实现波动率 Realized Volatility ACD模型 Autoregressive Conditional Duration Model |
| 英文摘要 | 随着全球经济的迅速发展，金融市场呈现出前所未有的波动，商业企业和金融机构都面临着日趋严重的金融市场风险，加强对金融市场风险的管理已经成为了金融机构和工商企业生存和发展的关键因素，而金融市场风险管理的核心是对风险的定量评估，即风险测度。因此，深入细致地探讨市场风险测度的方法，无疑具有一定的理论价值和实际意义。 目前我国的金融市场风险测度研究一般都采用低频日数据，这必然会损失部分日内信息，影响测度的准确性。一般而言，金融市场上的信息对金融资产价格变化的影响是个连续过程，离散模型必然会造成信息的丢失，数据频率越低，则信息丢失就越多。因此，考虑基于分钟、小时甚至秒、分笔等（超）高频数据计算金融市场风险，无疑为深化对金融市场微观结构的认识，提高金融风险测度的准确性提供了一个新的思路和方法。本文正是在这一思想的引导下，开展了基于高频数据和超高频数据的金融市场风险测度方法的研究。 文中以我国股票市场为研究对象，选取了上证指数、上证180指数、招商银行和贵州茅台四支股票进行实证分析，数据频率分别为分笔、5分钟、10分钟、15分钟、20分钟、30分钟和60分钟，时间跨度为2004-2010年。 本文旨在构建金融市场风险的高频数据计量体系，为使用高频或超高频数据测度金融市场风险的研究抛砖引玉，因此，本文在第一章对金融市场风险高频计量的相关文献进行了归纳总结，并在第二章对金融市场风险测度相关的金融市场理论基础及交易环境的变化情况进行梳理，指明了金融市场风险高频计量的研究方向，为该领域的研究向纵深方向发展打下了坚实基础；本文第三章对金融市场风险的一些基本概念进行了梳理，对金融市场风险测度的方法进行了综述，指出本文主要使用VaR方法来计算金融市场风险；使用高频数据测度金融风险的一个重要前提就是对金融高频数据的特征有很好的把握，因此，本文在第四章对我国股票市场上的高频数据特征和超高频数据特征进行了深入的研究；超高频数据的采集是随机间隔的，而高频数据是等间距采集的，数据特征的迥异决定了对两种类型的数据建立的模型不同，本文第五章针对我国股市上的高频数据特征，选用了已实现波动率模型对金融市场风险进行测度；第六章针对超高频数据的特征，选用超高频波动率模型对金融市场风险进行测度，选用蒙特卡洛模拟法对日内风险价值进行了测度，实证检验也验证了这些模型在我国金融市场上的适用性。本文研究的创新点可概括如下： 1．本文系统地梳理了基于高频数据的金融市场风险测度方法的相关研究，构建了金融市场风险测度的高频建模体系，对金融市场风险高频测度的进一步研究起到了抛砖引玉的作用。 2．比较深入地考察了高频数据和超高频数据的特征，数据特征的考察对于金融市场风险的测度建模至关重要，但以往研究经常忽略。 3．本文在测度方法的使用中，比较注重结合中国金融市场实际，对经典模型进行适当地改进及创新，建立了适合我国金融市场风险测度实际的高频数据模型体系。 4.本文依据统一的参照方法，对不同分布下的低频模型、高频模型和超高频模型测度方法的优劣性及适用性进行了评判，深入地研究了使用高频数据计算金融市场风险的精确性问题 。本文还对不同频率数据在测度金融市场风险时各自的适用性进行了研究和界定，这些在目前的研究中还不曾见到。; Along with rapid development of global economic, financial market is showing unprecedented volatility.Commercial companies and financial institutions are facing growing financial market risk, financial market risk management has become a core competence of financial institutions and even all enterprise's survival and development, and its core is quantitative assessment of risk, namely risk measurement. Therefore, the in-depth and detailed discussion of financial market risk measurement method has certain theoretical value and practical significance. At present, Financial market risk estimation research generally use the low-frequency data, which will inevitably lose some intraday information,and affect the accuracy of risk measurement. Generally speaking, financial market information is a continuous process of change on financial asset prices, the discrete model is bound to lose information, the lower frequency, the more missing information. Therefore, considering calculating financial market risk based on data whose frequency is minutes, hours or even seconds, tick-by-tick,have no doubt providing a new ideas and methods for deepening the understanding of the financial market structure, improving accuracy of the financial risk measurement .This article is in this thought, carried out the research of financial market risk measure methods based on high frequency data and uhf data. With Chinese stock market to be research object,we select the Shanghai Composite Index, SSE 180 Index, China Merchants Bank and Guizhou Maotai for empirical analysis. Data frequency contains tick-by-tick , 5m, 10m, 15m, 20m, 30m and 60m. The year spans from 2004 to 2010. This paper aims at building high-frequency financial market risk measurement system for better research of financial market risk measurement based on (ultra) high frequency data.Therefore, in the first chapter,we summarized related literature of financial market risk measurement based on high frequency data, and in the second chapter we summarized the financial market risk measurement theory and trading environment changes, pointed out the direction of the research of financial market risk measurement based on high-frequency data.In the third chapter of this paper ,we defined some basic concepts of financial risk, summarized the measure methods of market risk, and pointed out that this article mainly using VaR methods to calculate the financial market risk.An important prerequisite of financial risk measure using high frequency data is having a good grasp of the characteristics of high frequency data. Therefore, in the fourth chapter,we did a in-depth study of the characteristics of ultra high frequency data of the shanghai stock market.The ultra-high frequency Data acquisition is random intervals, and high-frequency data is collected equally spaced, the different data characteristics determines that data model are different. In the fifth chapter,according to high frequency data characteristics, we constructed realized volatility model on financial markets to measure risk.In the sixth chapter, according to ultra high frequency data characteristics, we constructed the UHFV models to measure the financial market risk, and constructed Monte Carlo simulation methods to carried out Intra-day value at risk. The empirical research verified the suitability of these methods. The innovation of this study can be summarized as follows: 1. This is the first systematic study of the frame of financial market risk measure methods based on high frequency data. We constructed the high-frequency modeling system of risk measurement, and played the role of catalyst in the research of financial market risk measurement based on high frequency data. 2. We did study on the the characteristics of high frequency data and ultra-high frequency data in depth. The study of the characteristics of the data is essential for modeling financial market risk, but previous studies often ignored. 3.This paper paid more attention to the improvement and innovation of classical model according to China's financial markets situation, and established a high frequency data model system which is suitable to China's financial market . 4. In accordance with the unified referencing methods, we assessed the inferiority and applicability of low-frequency model, high-frequency model and uhf model under different distribution, and did an in-depth researched on the calculating accuracy of the financial market risks based on high frequency data. The article also researched and defined the suitability of different frequency data in the measurement of financial markets risk, which has not yet seen in the current study .; 学位：经济学博士; 院系专业：经济学院计划统计系_统计学; 学号：15420080150259 |
| 语种 | zh_CN |
| 出处 | http://210.34.4.13:8080/lunwen/detail.asp?serial=29802 |
| 内容类型 | 学位论文 |
| 源URL | [http://dspace.xmu.edu.cn/handle/2288/38791]  |
| 专题 | 经济学院－学位论文 |
| 推荐引用方式 GB/T 7714 | 苗晓宇. 基于高频数据的金融市场风险测度研究, Measurement of Financial Market Risk Based on High Frequency Data[D]. 2011, 2011. |
| 个性服务 |
| 查看访问统计 |
| 相关权益政策 |
| 暂无数据 |
| 收藏/分享 |
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。
修改评论