中国股市噪声交易实证研究

发布时间：2018-01-01 18:22

本文关键词：中国股市噪声交易实证研究　出处：《哈尔滨工业大学》2013年博士论文　论文类型：学位论文

【摘要】：由于信息不对称，政府和投资者的博弈是一种独裁者博弈模式，有限理性的噪声交易者和非完全理性的噪声交易者占很大的比重，个人投资者和机构投资者都有从众心理，我国股市存在着传统金融理论（CAPM、EMH、MPT）无法解释的金融异象：由于投资者认知偏差造成的从众心理和“羊群效应”，投资者的反应过度和反应不足以及股票市场上的动量效应和长期反转效应等。本文从我国股市的实际出发，利用现代行为金融学有关噪声交易的理论，采用文献分析法、对比分析法和实证分析法对我国股市噪声交易的成分、杠杆效应和正反馈交易进行了实证研究。对噪声交易的理论进行了论述。包括噪声交易理论的由来，以及噪声交易的本质和内涵，并对噪声交易的风险和行为机制进行了界定。讨论了DSSW模型和噪声交易者的生成和获利机制。讨论了BSV、DHS和HS理论，以解释投资者的反应过度和反应不足，以及股票市场上的动量效应和长期反转效应。在分析中国股市噪声交易类型的基础上，，结合中国股市的发展态势分析了中国股市噪声交易的特殊表现：大幅波动、高换手率、受政策影响大等。研究了我国股市噪声交易的成分。以方差比检验（Variance-Ratio Test, VRT）为基础来构造一个噪声成分检验（Noise Composition Test, NCT）指标，并以此做纵向分析，考察中国股票市场在股改前和股改后的噪声成分。同时以近似全流通的股票和整个大盘做了横向比较。研究了我国股市噪声交易的杠杆效应。我国股票市场存在噪声交易的行为常常受到政策的干扰。而行业特征的区间受到杠杆效应影响不同。基于沪深300行业指数，利用ARMA-GARCH模型分析了沪深的各个代表性行业指数的波动性，GARCH（1,1）模型中各行业的a+均小于1且非常接近1，表明各个行业的GARCH过程是宽平稳的，行业的波动呈现集聚性和持续性，外部的冲击对条件方差的影响具有持久性。TARCH（1,1）和EGARCH（1,1）模型较好地说明各种外部因素对各行业的冲击影响，表现为杠杆效应和非对称性效应，且模型都消除了ARCH效应。这两个模型都说明了各行业间均存在明显的杠杆效应，“利坏消息”冲击比等量的“利好消息”冲击会产生更强的的波动。过度的噪声交易特别是正反馈交易是引起我国股市大起大落的主要原因。选用包括上证综合指数和深圳成分指数，中小板指数，沪深300指数，封闭式基金指数五个样本，采用非对称的GARCH-M模型，对我国股市存在的正反馈交易进行了实证研究。结果表明我国存在使用正反馈交易策略的噪声交易者，两市的日收益与噪声交易者行为有关；大盘蓝筹股的上市对治理噪声和遏制正反馈交易有积极作用；中小板的正反馈交易最为明显，其次是封闭式基金，然后是深圳成份指数、上证指数、最后是沪深300指数。利用资产定价模型BAPM，以受行政处罚的上市公司为样本研究过度的噪声交易。选用2007-2009年内受到中国证监会行政处罚的上市公司，从2006年1月到2009年12月的收益率作为分析的样本，同时取其对应的同行业的配对公司进行实证研究。发现噪声交易者风险（NTR）与股票超额收益是显著负相关的，显著性越高，那么投资者投资此类股票受到损失的可能性越大。本文在下述几个方面有所创新：第一，以VRT为基础构造NCT指标，分析了我国股市噪声交易的特殊性并进行了实证检验，并利用ARMA-GARCH模型对我国股市噪声交易引起的杠杆效应分行业进行了实证研究；第二，建立基于正反馈交易的证券资产收益率均衡模型，并结合非对称的GARCH-M模型对我国具有代表意义的五个指数进行了实证分析；第三，以行为资产定价模型BAPM为基础，分三种类型对受处罚的上市公司的噪声交易进行了的实证分析。
[Abstract]:Because of the information asymmetry, the game between government and investors is a dictator game mode, noise traders limited rational and non rational noise traders accounted for a large proportion of individual investors and institutional investors have a herd mentality, the stock market of our country exists in traditional financial theory (CAPM, EMH, MPT) cannot explain the financial anomalies because the herd mentality and "herding" caused investors cognitive bias, overreaction and underreaction, momentum effect on the stock market and long-term reversal effect. This article from the reality of our country stock market, the use of modern behavioral finance theory about noise trading, using literature analysis method, analysis method and composition an empirical analysis of noise trading in China's stock market comparison, leverage effect and positive feedback trading by empirical research.
The noise trading theory are discussed. Including the origin of the noise trading theory, and the essence and connotation of noise trading, and risk and behavior mechanism of noise trading are defined. The DSSW model is discussed and the generation of noise traders and the profit mechanism was discussed. BSV, DHS and HS theory, to explain the reaction of investors excessive and inadequate response, and the momentum effect on the stock market and long-term reversal effect. Based on the analysis of the types of noise trading on the stock market China, combining the development trend of the stock market analysis Chinese special table noise trading stock market: Chinese volatility, high turnover rate, affected by the policy and so on.
Study on the noise trading in China's stock market. By variance ratio test (Variance-Ratio Test VRT) as the basis to construct a noise component test (Noise Composition Test, NCT) index, this longitudinal analysis, the stock market in the study Chinese noise components before and after the reform. At the same time to do the comparison approximate full circulation of stocks and the whole market.
Study on the leverage effect of noise trading in China's stock market. China's stock market transactions are often affected by the presence of noise and interference. The policy of industry characteristics by interval leverage different effects. The CSI 300 industry index based on the analysis of the volatility of Shanghai and Shenzhen each representative industry index by the ARMA-GARCH model, the GARCH (1,1) the industry in the a+ model were less than 1 and is very close to 1, showed that the GARCH process of each industry wide is stable, the fluctuations of the industry has concentrated and persistent effects of external shocks to the conditional variance with persistent.TARCH (1,1) and EGARCH (1,1) model can better explain the impact of external factors on the the industry, as leverage effect and asymmetric effect, and the model can eliminate the ARCH effect. The two models are described both leverage effect exists between industries, "bad news" Chong The impact of "good news" than the same amount will produce stronger fluctuations.
Excessive noise trading especially positive feedback trading is the main cause of China's stock market. The change radically including Shanghai Composite Index and Shenzhen component index, small board index, the Shanghai and Shenzhen 300 Index Fund Index closed five samples, using asymmetric GARCH-M model, the existence of China's stock market positive feedback trading empirically the research results show that China's existing. Positive feedback noise traders trading strategy, the two cities daily returns and noise traders' behavior; blue chip listed on the noise control and containment of positive feedback has a positive effect; positive feedback trading of small plates is most obvious, followed by closed-end funds, then the Shenzhen component the index, the Shanghai index, the last is the CSI 300 index.
Use the capital asset pricing model BAPM in the administrative punishment of listed companies as a sample of excessive noise trading. The China Commission administrative punishment listed company is 2007-2009 years, from January 2006 to December 2009, the rate of return as the analysis sample, while taking the same industry should be matched companies to conduct empirical research. It is found that noise trader risk (NTR) and excess stock returns are negatively correlated, significantly higher, so investors investing in such stocks are more likely to suffer losses.
In this paper, some innovation in the following aspects: first, based on the VRT structure of NCT index, analyzes the particularity of noise trading in China's stock market and the empirical test, makes an empirical study on the leverage effect and the use of ARMA-GARCH model of noise trading in China's stock market sector; second, the establishment of rate equilibrium model of positive feedback trading the return of securities based on the index and the combination of five asymmetric GARCH-M model representative of China's empirical analysis; third, the behavioral asset pricing model based on BAPM, an empirical analysis is divided into three types of punishment of the listed company of noise trading.

【学位授予单位】：哈尔滨工业大学
【学位级别】：博士
【学位授予年份】：2013
【分类号】：F832.51;F224

【参考文献】