基于中国证券市场资产价格跳跃视角的市场交易行为、交易特征研究

发布时间：2018-03-05 21:01

本文选题：资产价格跳跃　切入点：交易行为　出处：《天津大学》2012年硕士论文　论文类型：学位论文

【摘要】：资产价格跳跃行为的识别、刻画及基于跳跃视角的市场交易行为、交易特征研究,能够深刻揭示资产价格的“发现—传导—融入—反馈”过程,使得投资者能够更为准确的把握资产价格行为的内在本质,进而为资产定价、资产配置和风险管理提供理论和实践依据,是目前证券市场微观结构研究的最前沿的领域之一。本文在金融市场微观结构理论视角下,基于BNS(2004,2006)提出的二次幂变差非参数跳跃识别经典框架,利用中国证券市场超高频交易数据,对中国证券市场日间低频跳跃以及日内高频跳跃进行了识别和刻画,并基于以上识别统计量研究了资产价格跳跃对市场交易行为的冲击过程,揭示了跳跃发生前后的市场交易特征变化。全文从四个方面进行了讨论和研究: 第一部分简要阐述了本文的研究背景,结合金融市场微观结构与资产价格行为建模理论背景与我国证券市场目前存在的问题提出了研究问题及研究意义,在此基础上介绍了资产价格跳跃识别研究进展,并对资产价格跳跃与市场交易行为、市场交易特征相关研究现状进行了综述,为下面的研究内容做了理论铺垫。第二部分首先对资产价格跳跃识别方法进行了详细阐述,分别介绍了资产价格跳跃参数及非参数识别方法体系,在此基础上以资产价格跳跃非参数体系为研究基础,详述了目前两种较为经典的非参数跳跃识别方法。最后,在以上跳跃识别的理论基础上,基于BNS提出的二次幂变差理论跳跃识别框架,采用中国证券市场上证综指高频交易数据,对我国证券市场的日间/日内高频跳跃进行了识别和描述,研究发现我国证券市场资产价格跳跃行为发生较为频繁,其中存在大量日内连续跳跃现象,且对于不同的个股来说,跳跃的频率及幅度特征呈现出较大差异性,而大多数跳跃行为发生在开盘时间段。第三部分将中国证券市场个股日内高频跳跃发生时刻作为虚拟变量,对MRR模型进行了扩展和改进,通过Jump-MRR模型研究了资产价格跳跃对于市场交易行为的影响过程。研究发现,资产价格跳跃对交易行为存在显著的冲击影响,跳跃发生前,平均信息冲击成本和流动性成本均低于非跳跃时刻;跳跃发生时,信息冲击成本明显冲高,表明跳跃是一种非常剧烈的信息融入价格的现象,同时跳跃发生时流动性成本显著降低,表明跳跃的发生伴随着投资者的交易意愿的增加;跳跃发生之后5分钟左右,信息冲击成本和流动性冲击成本均存在一个迅速回复过程。最后,资产价格跳跃发生时不同市值规模的股票对应的交易行为呈现出显著的差异,随着公司市值规模增大,发生跳跃时股票平均信息冲击成本和流动性成本变小,显示出资产价格跳跃发生时高市值股票流动性较好、信息非对称程度较低,投资者对该类股票的交易意愿强烈。第四部分为资产价格跳跃与市场交易特征关系研究。该部分刻画了资产价格跳跃前后的交易量特征变化模式,并结合极端市场下的投资者情绪,研究了资产价格跳跃与市场交易量、收益率特征之间的关系,从更为微观的角度揭示信息快速融入过程中投资者的反应及市场交易状态和影响因素。研究发现,我国证券市场在资产价格跳跃发生时往往伴随着交易量的大幅增加,跳跃幅度与即刻交易量之间显著正相关,跳跃发生后交易量存在迅速下降的过程;正向、负向跳跃所代表的利好、利空的市场冲击对市场交易量、收益率存在非对称影响,整体来说,投资者对利空消息更敏感,但在不同的市场情绪下投资者对市场冲击存在非理性反应,牛市时投资者倾向于对利空消息反应不足,而在熊市时倾向于对利好消息反应不足。
[Abstract]:The identification of asset price jump behavior, and describe the market transaction behavior jump based on the perspective of the characteristics of the transaction, can reveal the asset price discovery - transmission into the feedback process, in essence allows investors to more accurately grasp the asset price behavior, and asset pricing, provide theoretical and practical basis asset allocation and risk management, is currently one of the most cutting-edge research on the micro structure of the securities market in the field.
This article in the micro financial market structure from the perspective of the theory, based on the BNS (20042006) proposed two power variation of non parametric jump identification classic framework, ultra high frequency trading data by China securities market, the securities market of Chinese day low frequency hopping and high-frequency intraday jump for the identification and characterization, and based on the above statistics of asset recognition the price jump on the market trading behavior impact process, reveals the market trading characteristics before and after the jump change. This paper from the four aspects of the discussion and research:
The first part briefly introduces the research background of this paper, combined with the existing background of financial market microstructure and asset price behavior modeling theory and problems of China's stock market, put forward the research questions and research significance, introduced on the basis of the progress of asset prices jump identification, and asset price jumps and market transactions, are reviewed market characteristics of relevant research, as the theoretical foundation for the following research.
The second part elaborates on the asset price jumps identification method are introduced. The asset prices jump system parameter and non parameter recognition method based on non parametric asset price jumps system as the research foundation, described the current two kinds of more classical non parametric jump identification method. Finally, the above theoretical basis of jump recognition on the variational theory of jump recognition framework is put forward based on a power of two BNS, the Chinese Shanghai stock market high-frequency data on China's securities market day / days of jump by the recognition and description of asset price discovery in China's securities market jumps occur more frequently, there are a lot of day in the continuous jumping phenomenon, and for different stocks, frequency and amplitude characteristics of jumping shows a large difference, and most jump behavior occurred in the opening Time slot.
The third part will Chinese stock market intraday jump moment as dummy variables and the MRR model was extended and improved, through the Jump-MRR model to study the process of asset prices jump effect on the market trading behavior. The study found that asset prices jump to the transaction there was a significant impact, jumping, average the impact of information costs and liquidity costs were lower than non jumping moment; jump occurs, the information costs shot up significantly, indicating that the jump is a very severe information into the price of the phenomenon, at the same time jump occurs when the liquidity cost is significantly reduced, indicating that the jump occurred with the increase in investors willingness to trade after the jump; about 5 minutes, showed a rapid response process the information costs and liquidity impact cost. Finally, asset price jumps occur with market value The size of the corresponding stock trading behavior showed significant differences, with the company's market capitalization size increases, the jump average stock information impact cost and liquidity cost decreases, showing the asset price jumps when the high market value of the stock liquidity is better, with lower degree of asymmetric information, investor of the stocks trading wishes.
The fourth part is the research on the relationship between asset price jumps and market transaction characteristics. This part describes the change pattern of asset price jumps and trading volume characteristics, and combined with the extreme market under the investor sentiment, of asset prices jump and market trading volume, the relationship between yield characteristics, reveal the factors from rapid integration into the information process of investors the transaction status and market response and influence more microcosmic view. The study found that China's securities market in asset price jumps is often accompanied by a substantial increase in transaction volume, a significant positive correlation between the jump and immediate trading volume, trading volume after the jump process decreased rapidly; the positive, negative jump on behalf of the good, the bad market impact on the market trading volume, there is asymmetric effect, the overall rate of return, investors are more sensitive to bad news, but in different In market sentiment, investors have irrational responses to market shocks. When bull market, investors tend to react poorly to bad news, but tend to react poorly to good news in bear market.

【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】：F832.51;F224

【参考文献】