【作者】 张小勇；

【导师】 马超群；

【作者基本信息】 湖南大学 ， 工商管理， 2013， 博士

【摘要】 传统的金融理论基于有效市场假说,以“价格可以充分反映该时点所有可得信息”为前提,仅关注金融市场中资产价格的时间序列特征,仅从价格本身出发来对价格波动进行解释和预测。然而,价格波动的复杂性让学术界开始对这一前提产生了怀疑,交易量---这个被忽视的可能包含市场信息的因素,随着金融市场微观结构理论的发展而逐渐受到学者们的重视。其实在投资界,“量价结合”这一准则早已被人们熟练应用在金融市场的技术分析上。因此,从理论与实证的角度对金融市场的量价关系进行深入研究显得尤为必要。本文基于混合分布假说研究金融市场中的量价关系以及量价关系背后的主要驱动因子。研究从两个方面展开,在股票市场与期货市场分别利用基于低频数据的GARCH-V模型检验与基于高频数据“已实现”波动率线性模型检验的方法,来横向比较研究交易量对市场价格波动的解释能力、纵向方法创新研究去异方差交易量与价格波动的关系,以及深入挖掘中国期指市场的量价规律与产生量价关系的主要驱动因子。具体内容如下：首先,从不同的GARCH族模型以及非正态GARCH族模型出发,多方位的比较研究了中国股市的价格波动特征,并对市场风险进行了VAR度量。研究表明,EGARCH模型和APARCH模型的效果优于其他模型,且学生t分布假设和GED分布假设下的GARCH族模型在总体上要优于正态分布假设,这为今后在针对中国股市选择波动性模型时,提供了重要的参考价值。其次,在价格波动方程中加入交易量,利用基于GARCH-V模型的实证检验方法,横向比较研究了七个国家股票指数交易量对市场价格波动的解释能力；接下来,剔除掉交易量序列的波动丛聚性,创新性的研究了去异方差交易量与价格波动的关系。研究表明,成熟市场上交易量对价格波动的解释能力相对较强,市场价格对信息的吸收和反映能力较强,市场的有效程度较高；并且,去异方差交易量是更好的信息流代表,能够增加交易量对价格波动的解释能力,市场成熟度越高的国家,去异方差交易量对价格波动的解释能力越强。随后,以沪深300股指期货为研究对象,根据Jone等(1994)的研究将成交量划分为成交次数和平均交易头寸,并考虑“已实现”波动率的跳跃和非对称性特征,构造了中国股指期货市场量价关系的基础模型、连续和跳跃波动模型及量价关系非对称模型。研究表明：沪深300股指期货的成交量与价格波动之间呈现明显的正相关关系；成交量、成交次数及平均交易头寸对连续和跳跃波动都有显著的正向影响,且成交量与连续波动的正相关关系可以较为精确的反映我国期指市场总的量价关系；下偏已实现半方差较上偏已实现半方差包含更多的市场波动信息；平均交易头寸作为量价关系背后的主要驱动因子,可以更好地解释市场波动。接下来,用GARCH-Copula模型研究股市量价尾部关系,这不仅考察了价格高涨与高交易量,价格大跌与低交易量之间的关系,还考察了大的价格变动与高交易量、小的价格变动与低交易量之间关系。这刻画了在极端市场条件下,量价间尾部的相依性,同时具有时变的特征。最后,总结了本文的主要研究成果与创新点,提出未来研究方向与展望。更多还原

【Abstract】 Traditional financial theory based on the Efficient Market Hypothesis, which, assuming that price can fully reflect all information available at that time, only focuses on the time series characteristics of the asset price, and strives to explain and predict the price volatilities only with the price itself. Due to the complexity of the price volatility, the academia, however, has called the premise into question; with the development of the micro-structure of the financial market, volatility--the neglected factor which might possibly involve market information--has gradually come under scrutiny. Actually, the thumb rule of the "volume-price analysis" has long been applied adeptly by practitioners in technical analysis in the financial market. Therefore, it is particularly necessary to explore the possible relationship between the trading volume and the asset price theoretically and empirically.This paper studies the price-volume relation in financial markets and the principal driving factors behind it on tha basis of the Mixture Distribution Hypothesis (MDH). This study unfolds from two aspects:1) in stock market, we compare the explanatory power of trading volumes to price volatility in different countries’stock markets and develop a new method to deal with the trading volume as persistence-free series using a GARCH-V model and low-frequency data;2) in the futures market, we deeply dig the principal driving factors for the price-volume relationship in the Chinese stock index futures market using some linear models with the "realized volatility" and high-frequency data. The specific contents are as follows:First, starting from different GARCH-type models as well as non-normal GARCH-type models, the characteristics of the price volatility of China’s stock market is studied using some GARCH-type models, and a VAR measure of market risk is developed. The study shows that the EGARCH model and APARCH model perform better than other models, and the GARCH-type models under the assumption of the student t distribution or the GED distribution, in general, work better than GARCH models under the assumption of the normal distribution. The findings serve as an important reference for the selection of volatility models to research in China’s stock market.Subsequently, with the introduction of trading volume into the price volatility equation, the GARCH-V model is employed to empirically test the explanatory power of trading volume to market price volatility in seven countries’stock market. The volatility clustering of the trading volume is removed and the relationship between the persistence-free trading volume and then the price volatility is innovatively explored. The findings demonstrate that the trading volumes of mature financial markets have more explanatory power to price volatilities, and market prices assimilate and reflect information better, which implies high efficiency of the financial market. More importantly, as the explanatory power of the persistence-free trading volume to the price volatility increases with the maturity of the countries’financial markets, the persistence-free trading volume is a more desirable proxy for the information flow, and capable of enhancing the explanatory power of the trading volume for the price volatility.Next, according to Jone et al.(1994), we divide its trading volume of the CSI300stock index futures into trading times and average trading sizes, and take the jumps and the non-symmetries in the "realized" volatility into account to construct a base model, a continuous and jump volatility model and a non-symmetrical model for volume-price relationship models for China’s stock index futures. The study shows that there is an significantly positive correlation between the trading volume and the price volatility; the trading volume, the trading times, and the average trading sizes all have significantly positive effect on the continuous and the jump volatility; the positive correlation between the trading volume and the continuous volatility can reflect more accurately the aggregate volume-price relation in China’s futures index market; the downside realized semi-variance includes more market volatility information than the upside realized semi-variance; and the average trading size, as a major driving factor behind the volume-price relationship, has more explanatory power for market volatility.Then, the relationship between the volume and the price at the tail is studied using a GARCH-copula model. The model not only examines the relationship between high price rises and high volumes, and between big price falls and low volumes, but also the relationship between big price changes and high volumes, and between small price changes and low volumes. This characterizes the interdependence between price and volume at the tail under extreme market conditions, and meanwhile shows the time-varying features.Finally, we summarize the main findings and contributions, and propose directions in future research.更多还原