【作者】 王晓琳；

【导师】 叶强；

【作者基本信息】 哈尔滨工业大学 ， 管理科学与工程， 2016， 博士

【摘要】 关于投资者行为如何影响金融市场,传统金融学并没有给出答案,但在行为金融学框架下,投资者行为是影响证券价格和市场运行的重要因素。大数据的飞速增长正在影响人类日常生活的各个方面,大数据的出现为我们解决人类赖以生存复杂世界的基本问题提供了一个重要的新机会,金融市场是此类定量调查的主要目标。在互联网迅速发展之前,理论界通常利用股票交易量、报纸头条新闻以及广告费用等作为考察投资者的代理变量,谷歌搜索趋势这一用户创造内容平台的出现为投资者提供了便捷的信息获取方式,互联网环境推动了投资者行为与金融市场之间关系的深入研究。系统地研究投资者在金融市场上的行为及其产生的影响,能够使投资者更清楚自己在预测股票价格时出现了怎样的偏差,并可能有助于投资者依据这些偏差获取超额收益;有助于更好的理解收益、波动性和交易量之间的互动机理;有利于为监管机构更好地掌握投资者的心理特征,从而为其更有针对性地调控市场提供理论基础。本研究主要内容如下:首先,本文运用标准普尔500股票的谷歌搜索趋势记录这一用户创造内容的平台作为投资者关注的代理变量,探讨投资者关注对分析师盈余预测行为的影响,以及分析师预测行为对谷歌搜索趋势的影响。实证结果发现搜索量与分析师预测修正、分析师总人数和分析师预测分歧相关。与之前的研究结论一致,我们发现分析师盈余预测分歧较大的股票其预期收益将显著下降,即分析师盈余预测分歧效应。然而,当投资者关注度较高时,这种分析师盈余预测分歧效应被显著削弱,研究结果为传统金融学无法阐明的金融异象提供了理论依据。其次,本文首次发现我国股指期货市场流动性日内效应与周内效应,得出了期货市场流动性的日内模式在上午呈现反“J”形变化模式,下午呈现倒“U”型变化模式,期货流动性的周内模式呈现倒“V”形变化模式。实证结果表明中国股票指数期货市场流动性的周内模式和日内模式存在着显著的周期性。本文丰富了对股票指数期货市场流动性的订单驱动型市场的研究,所获得的实证结果有助于股票指数期货投资者更好地了解市场微观结构的投资者行为。再次,采用股指期货高频数据对中国股指期货运行初期期货与现货市场的领先滞后关系进行研究,分别在投资者情绪高涨、投资者情绪低落与投资者情绪平稳三个区间进行实证分析,探讨股指期货价格发现功能在投资者情绪表现不同时是否均能得以实现。本文研究结果修正了之前关于股指期货不具有价格发现功能的结论,股指期货价格发现功能的确立,有助于我国金融市场的完善。第四,基于指令失衡这一关键指标作为投资者交易行为的代理变量,分析投资者交易行为与价格行为之间的关系,具体包括投资者交易行为的影响因素,投资者交易行为与市场流动性变化关系,投资者交易行为对市场波动性与成交量关系的影响。研究结果发现在中国农产品期货市场指令失衡具有收益预测的能力。过度指令失衡引起的价格行为并不符合传统库存理论的解释,中国期货市场的价格行为主要由投机者的羊群效应导致。本文运用多种统计学与计量经济学中常用的分析方法,包括:时间序列回归分析、协整检验、VECM模型估计、格兰杰因果检验与Fama-Mac Beth截面回归模型等,具有前沿性、系统性、市场性和监管导向性。更多还原

【Abstract】 The traditional financial theories did not give the answer that how investor behavior affect the financial market. However, under the framework of behavioral finance, investor behavior is an important factor affect s the prices and the operation of financial market. Accordingly, the researches of this subject have many contents. Systematic study of investors behavior and the influence of investors in the financial market make investors more clearly their bias in forecasting stock price and may be useful to investors obtain excess returns with these bias. As well as the research of investor behavior will contribute to better understanding of return, help investor in deep understanding the dynamic interaction mechanism between return and trading volume, volatility. Further, it is also beneficial for regulators to better grasp the psychological characteristics of investors. Provide regulators a theoretical basis for their more efficiency control the market.The contents of this paper are as follows:Firstly, using Google trends records for S&P 500 stocks, we explore whether analysts’ forecasts influence investor attention, Google search trends, a typical user-generated content platform and whether Google search trends have an impact on the effects of analysts’ forecasts on future stock returns. We find that search volume is related to analyst earning forecast updates, percentage of negative forecasts, total number of forecasts and forecast dispersion. Consistent with previous studies, we find that stocks with higher analysts’ forecast dispersion earn significantly lower returns than otherwise similar stocks. However this dispersion effect disappears if Google abnormal search volume is high. Our evidence suggests that information asymmetry associated with analysts’ forecast dispersion impacts both investor attention and the dispersion effect on stock returns.Secondly, this paper examines weekday and intra-day liquidity effect of CSI300 Stock Index Futures. Empirical results show that weekday liquidity pattern exhibits a reverse V-shaped for the CSI300 index futures, while the intra-day liquidity pattern exhibits a reverse J-curve in morning and a reverse U-curve in afternoon. Furthermore, this paper investigates the influencing factors of stock index futures liquidity and gives some recommendations for the trading regime of Chinese Stock Index Futures market.Thirdly, this paper presents price index and index futures lead lag relationship for empirical analysis based on high-frequency data during three different periods, high investor sentiment, low investor sentiment and flat investor sentiment respectively. Using the unit root test, cointegration test, VECM estimates, Granger causality test method, we explored the lead lag relationship between CSI300 spot index and index futures with 5 minutes high-frequency data. The empirical results show that with the increase of transaction time, stock index futures in a larger extent guiding spot index futures, price discovery function of stock index futures market are better.Fourth, due to the volume is often divided into a small sum of investors, and the volume itself can not show the direction of a transaction, order imbalance is better to reflect trading behavior. This paper selects nine futures contracts from Dalian Commodity Exchange(DCE) and Zhengzhou Commodity Exchange(ZCE) as data example to study trading activity and the price behavior based on the high frequency data from 2010 to 2015. In this paper, we find that contemporaneous order imbalances are positively related to returns. Order imbalanc e caused by price pressure last more than one day indicating difficult y in absorbing excess buy and sell orders. We also find that lagged order imbalance can predict current returns and that the effect of order imbalance on liquidity is limited. These results are consistent with the explanation that speculative trading and herd effect not liquidity hinders the Chinese agricultural futures markets to accommodate excess order imbalance.In this paper, we used a variety of statistical and econometric analysis method, including: time series regression analysis, cointegration test, VECM model estimation, Granger causality test and the Fama-Macbeth cross-sectional regression model. The main characteristics of this paper are regard of leading, system, market and guiding for regulatories.更多还原