基于MEM模型的我国股指期货市场波动率研究

发布时间：2018-06-14 18:14

本文选题：高频数据 + 乘积误差模型（MEM）　；参考：《天津大学》2012年硕士论文

【摘要】：2008年底爆发的全球性金融危机对世界经济造成的影响至今仍未平息，此次危机凸显了金融风险监管上的巨大漏洞，引发了国际社会对于金融创新与金融监管的大讨论；2010年4月16日，我国正式推出了股指期货交易，其在丰富了我国资本市场交易品种的同时，也对我国尚未完善的监管体制带来了新的难题。金融高频数据计量经济学自20世纪90年代以来成为了金融计量学、金融工程学的一个重要研究领域，高频数据的波动建模作为其中一个重要的研究方向，旨在通过适当的模型深入分析金融市场的波动特征。乘积误差模型（MultiplicativeErrorModel,MEM）由Engle（2002）提出，其适用于对于高频非负（non-negative）金融数据进行建模。本文即基于MEM模型对于中国股指期货市场的波动进行建模与研究。本文的主要研究工作和创新点如下： 1、虽然MEM模型被国外学者广泛用于实证分析，但我国学者在这方面所做的研究极少，本文利用MEM模型对我国股指期货合约价格指数的“已实现”波动建模，并在其中考虑了随机误差项可能出现的不同分布函数的情形，分析不同随机误差项分布对建模结果的影响。 2、Anderson和Bollerslev(1998)提出了“已实现”波动（RealizedVolatility，RV）这一崭新的波动度量方法，基于其计算简单，无需建模、进行参数估计等优良特性，受到了学者们的普遍关注。虽然徐正国（2004）提出了调整“已实现”波动（Adjusted Realized Volatility）、郭名媛（2006）提出了赋权“已实现”波动（Weighted Realized Volatility），并从理论和实证两个方面均证明了自己的改进形式比“已实现”波动性质更优，但国内尚无学者利用实证研究比较这三者的统计性质，也没有学者利用同一种模型为这三种“已实现”波动建模。本文对于股指期货合约价格指数的“已实现”波动、调整“已实现”波动和赋权“已实现”波动从统计特性和MEM建模结果这两方面进行了比较与研究。 3、本文基于门限ACD模型的思想，，建立门限MEM模型（ThresholdMultiplicative Error Model），并利用其对我国股指期货合约价格指数的“已实现”波动建模。本论文是国家自然科学基金项目《基于MEM模型的金融市场分析》（No：70901055）的组成部分。
[Abstract]:The global financial crisis that broke out at the end of 2008 has not yet calmed down the impact of the global financial crisis on the world economy. The crisis has highlighted the huge loopholes in the supervision of financial risks and triggered a great discussion on financial innovation and financial regulation in the international community. On April 16, 2010, China officially launched stock index futures trading, which not only enriches the trading varieties of China's capital market, but also brings new difficulties to China's imperfect regulatory system. Financial high-frequency data econometrics has become an important research field of financial metrology and financial engineering since 1990s. Volatility modeling of high-frequency data is one of the important research directions. The purpose of this paper is to analyze the volatility characteristics of financial markets through appropriate models. The product error model / Multiplicative error Model (MEM) was proposed by Engle2002.It is suitable for modeling high-frequency non-negative financial data. Based on the MEM model, this paper models and studies the volatility of China's stock index futures market. The main research work and innovations of this paper are as follows: 1. Although MEM model has been widely used in empirical analysis by foreign scholars, Chinese scholars have done very little research in this area. In this paper, we use the MEM model to model the "realized" fluctuation of the price index of stock index futures contracts in China, and consider the different distribution functions of the stochastic error term. This paper analyzes the influence of different random error term distribution on the modeling results. (2) A new wave measurement method "realized" is proposed by Anderson and Bollerslev (1998). Based on its simple calculation, no need for modeling and parameter estimation, a new wave measurement method is proposed. It has received widespread attention from scholars. Although Xu Zhengguo (2004) proposed to adjust the "realized" volatility, Guo Mingyuan (2006) proposed that the "realized" wave has been realized, and proved that its improved form is better than the "realized" volatility in both theory and practice. However, no domestic scholars use empirical research to compare the statistical properties of the three, and no scholars use the same model to model the three "realized" fluctuations. This paper deals with the "realized" fluctuation of the price index of the stock index futures contract. Adjusting "realized" and "weighted" realized "volatility is compared and studied in terms of statistical characteristics and MEM modeling results. 3. This paper is based on the idea of threshold ACD model. The threshold MEM model is established to model the realized volatility of China's stock index futures contract price index by using the threshold MEM model of Threshold-Multiplicative error Modeler. This thesis is a part of National Natural Science Foundation Project "Financial Market Analysis based on MEM Model" No: 70901055.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F832.5

【参考文献】