【作者】 李元；

【导师】 张谊浩；

【作者基本信息】 南京大学 ， 金融学， 2014， 硕士

【摘要】 本文运用在中国市场占有率最高的搜索引擎百度搜索的数据,围绕着网络搜索能否预测中国的股票市场这个问题展开研究。本文认为投资者网络搜索是一个复合了投资者情绪、投资者关注以及其他一些因素的灰箱,并且相对于传统的投资者情绪和投资者关注指标,网络搜索对股票市场的解释力和预测力可能更强,对此本文构建了一个理论模型进行了详细阐述。在实证部分,本文运用2006年6月到2012年5月沪深300指数上市公司数据和百度搜索指数数据,研究发现：(1)投资者网络搜索强度对股票短期收益率、短期交易量及累积收益率均有影响,股票短期收益率也影响投资者的网络搜索,表明网络搜索与股票市场具有相关性。(2)股票市场能影响网络搜索,但网络搜索可以在更大程度上影响并预测股票市场的表现,网络搜索与股票收益间存在的内生性问题对预测效果的影响很小。根据网络搜索构建投资组合可以获取超额收益。(3)网络搜索强度指标与传统的投资者情绪和投资者关注指标间存在较强的相关关系；相较于投资者情绪和投资者关注,投资者网络搜索对股票市场的解释力及预测效力更强。(4)在股指上升阶段,网络搜索对股票市场收益基本没有解释力和预测力,而下降及调整阶段,网络搜索可以很好的解释和预测股票市场。本文的研究有着比较大的意义：在学术方面从理论和实证拓展了现有的研究；在现实方面,无论对个体投资者还是对机构投资者,本文设计的投资策略以及不同阶段网络搜索对股票市场预测效力的分析比较都有着重要的参考价值。而对于证券监管部门而言,若能挖掘、捕捉并监测好投资者网络搜索指标的话,则能够改进其股票市场监管的技术手段,优化市场监管绩效。更多还原

【Abstract】 Using Baidu search engine data, this paper tries to answer this key problem:Can internet predict China’s stock market. We argue that investors’ internet search is a composite of investor sentiment, investor attention and other factors, which has a stronger power to explain stock market compared to conventional indicators of investor sentiment and investor attention. A theoretical model is constructed to elaborate this viewpoint.In this paper, the highest market share of Chinese search engine Baidu search data around the Internet search can predict China’s stock market, a study of this issue. This paper argues that investors in Web search is a composite of investor sentiment, investors are concerned about a number of other factors as well as gray boxes, and compared to conventional indicators of investor sentiment and investor concerns, the stock market Web search explanatory power and predictive power may be stronger, this paper constructs a theoretical model in detail.Based on CSI-300 index constituent stocks and Baidu search index from 2006.6 to 2012.5, we find:(1) Internet search intensity can predict the stock’s short-run return, trading volume and cumulative return, which implies a positive relationship between internet search and stock market.(2) Internet search intensity can affect stock market further and predict stock market effectively. The problem of endogenous has little impact on the prediction effect. Moreover, a simulation analysis reveals that a stock portfolio constructed according to the internet search will yield excess abnormal return.(3) Internet search intensity index is strongly correlated with traditional investor sentiment index and investor attention index, and is more powerful to explain and predict stock market.(4) Internet search intensity index has little power to explain and predict stock market when stock index is in rising phase, but has strong power when in decline and adjustment phase.This paper expands the existing research in academic theory and empirical analysis. For individual investors and institutional investors, the investment strategy and the comparative analysis of different phase have important reference value. For the securities regulatory authorities, it is possible to improve the technical means of stock market regulation and supervision.更多还原