基于高频数据的投资者交易行为研究

发布时间：2018-04-13 16:04

本文选题：投资者行为 + 高频数据　；参考：《首都经济贸易大学》2014年博士论文

【摘要】：伴随着证券市场的蓬勃发展以及信息技术、数据处理技术的飞速进步，金融数据由原来月、周以及日数据逐步细到小时、分钟甚至秒的数据，这种超高频数据的采集对于我们金融领域研究，尤其是当今行为金融研究盛行的时代，显得异常重要。行为金融将心理学和金融学结合起来进行研究，认为投资者的决策不仅受到客观市场的制约，还会受到来自自身或周围群体的主观方面的影响，因而投资者会表现出各种异常，在证券市场上相应地体现为各种异象，对于这些异象的解释也就层出不穷。本文针对人们聚焦的投资者行为，从行为金融理论出发，运用适合（超）高频数据的自回归条件持续期模型（ACD model），通过对它基本形式的合理扩展，得到了相应的不同形式。具体针对市场微观结构、事件冲击以及投资者情绪几个方面，不仅做了理论分析，而且结合沪深300样本股中的平安银行股票分笔数据一一做了实证分析。首先，通过构建价格持续期，在基本ACD模型中引入三个代表市场微观结构的变量——买卖价差、每笔平均交易量和交易密度，，分别和同时分析了它们与持续期（久期）之间的关系。发现三者与久期均具有显著的负相关关系，并根据微观结构理论对此进行了解释，认为这是由于信息交易者的存在导致，而非流动性造成。其次，通过构建交易量持续期，在基本ACD模型中考虑了非对称效应，分析了事件冲击对投资者行为的影响。发现在预期事件冲击下，非对称效应并不大，但在非预期事件（以2013年8月16日光大事件为例）冲击下，非对称效应则较大，据此得到了投资者行为针对事件冲击的前后反应过程。最后，通过对衡量投资者情绪指标的分析，选择了换手率作为分笔数据中代表投资者情绪的变量，并且买卖方向分别代表了投资者的乐观和悲观情绪，实证发现，无论是收益率，还是波动性，都与投资者情绪呈现正相关关系。通过理论及实证研究，进而提出一些对策建议，即投资者和政府通过对持续期的观察，在面对不同大小的久期时，也即面临不同大小的风险时，应该改变交易频率，并且采取措施规避风险。另外，为了有效解决市场流动性问题，减少甚至消除信息交易带来的噪声影响，市场监管者应该着力改善市场机制，让信息更为透明，有效解决信息不对称问题，唯有这样，才能使我国证券市场更好更快地发展。
[Abstract]:With the rapid development of the securities market and information technology, data processing technology, financial data from the original month, week and day data gradually fine to hours, minutes or even seconds of data,The acquisition of UHF data is very important for our financial research, especially in the era of behavioral finance.Behavioral finance combines psychology with finance. It is believed that investors' decisions are not only restricted by the objective market, but also influenced by the subjective aspects from themselves or the surrounding groups, so the investors will show various anomalies.In the securities market, they are reflected in various visions, and the explanations of these anomalies emerge endlessly.In this paper, based on the theory of behavioral finance, the autoregressive conditional duration model (ACD model) suitable for ultra-high frequency data is applied to the focused investor behavior. Through the reasonable expansion of its basic form, the corresponding different forms are obtained.Aiming at the market microstructure, event impact and investor sentiment, this paper not only makes a theoretical analysis, but also makes an empirical analysis based on the data of Ping an Bank in Shanghai and Shenzhen 300 sample stocks.First of all, by constructing the price duration, we introduce into the basic ACD model the variation of the market microstructure of the three represents-the price spread, the average transaction volume and the transaction density per transaction.The relationship between them and duration (duration) was analyzed separately and simultaneously.It is found that there is a significant negative correlation between the three factors and duration, which is explained by the theory of microstructure, which is due to the existence of information traders rather than liquidity.Secondly, the asymmetric effect is considered in the basic ACD model, and the impact of event shock on investor behavior is analyzed by constructing the duration of trading volume.It is found that the asymmetric effect is not large under the shock of the expected event, but the asymmetric effect is larger under the shock of the unexpected event (as in the case of the Everbright event on August 16, 2013).Based on this, the reaction process of investor behavior in response to event shock is obtained.Finally, through the analysis of the measure of investor sentiment, the paper selects turnover rate as the variable representing investor sentiment in the split data, and the direction of trading represents the optimism and pessimism of investors respectively.Both yield and volatility are positively correlated with investor sentiment.Through theoretical and empirical research, this paper puts forward some countermeasures and suggestions, that is, investors and governments should change the trading frequency when facing different length of time, that is, risk of different size, by observing the duration of the period.And take measures to avoid risk.In addition, in order to effectively solve the market liquidity problem and reduce or even eliminate the impact of noise caused by information transactions, market regulators should focus on improving the market mechanism, making information more transparent, and effectively solving the problem of information asymmetry. Only in this way,Only then can our country's securities market develop better and faster.
【学位授予单位】：首都经济贸易大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：F224;F832.51

【引证文献】