市场公开信息对我国股票收益率异质性影响的实证研究

发布时间：2018-01-23 20:09

本文关键词： 股票收益率市场公开信息贝叶斯分位数模型投资心理偏差　出处：《上海大学》2013年硕士论文　论文类型：学位论文

【摘要】：股票收益率是否能够反映市场信息及其反映程度一直是理论界和实务界较为关心的问题。在理论上，最早由Fama（1973,1992）提出市场有效性理论和三因子模型来分析股票收益率对于市场信息的反应程度和影响因素。在实务界，投资者通过基本面和技术面信息来对股票收益率进行分析，借以取得盈利。近年来随着行为金融学的兴起，以De Bondt（1985）为代表的行为金融学学者从投资者心理的角度来分析投资者对于市场信息是否存在过度反应和反应不足，从而论证股票收益率对于市场信息可能存在的异质性反应。基于Fama和De Bondt对于股票收益率反应特征的相关理论，本文主要利用贝叶斯统计模型来实证分析我国股票收益率对于公开市场信息的反应及其反应程度。我们首先将市场公开信息分为3类，即企业财务信息、股票市场信息和宏观经济信息，并通过AIC和贝叶斯因子等模型选择标准来选择较为显著的变量来进行回归。实证结果显示相比企业财务信息，股票市场信息和宏观经济信息对于股票收益率有更大的影响作用，其中股票换手率、市场风险系数β，存款准备金率和货币供给量对于股票收益率的影响作用非常显著，这说明我国的投资者在进行投资决策时，比较重视股票市场技术面信息和宏观信息。我们认为这主要是由于我国股票投资者短期投机、股票市场分红制度、财务信息披露不完善所致。对于实证我国股票投资者是否存在过度反应和反应不足方面，以往学者们主要通过构筑“赢者组合”和“输者组合”，并比较投资收益率进行实证分析。在本文中，我们通过贝叶斯分位数回归来实证分析我国投资者对于市场信息是否存在过度反应和反应不足。在估计过程中，，我们使用马尔可夫链蒙特卡罗模拟（MCMC）方法并通过Gibbs进行抽样。研究结果表明对于那些收益率出现大幅度上升的股票，投资者倾向于高估未来股票上涨的可能并夸大企业规模、换手率和货币供应量等利好消息，这使得股票对于一个单位利好信息的冲击，股票收益率的上升程度要大于其平均的上升程度。而对于那些出现暴跌的股票，投资者对上述忽视利好消息的作用，这使得投资者对于一个单位利好信息的冲击，股票收益率的上升程度要远低于其平均的上升程度。所以我们认为我国投资者也存在行为金融学中所提到的过度反应和反应不足。究其原因，我们认为这种现象主要是我国投资者的自我归因偏差、保守性心态等投资心理偏差所致。为了使得我国股票市场更有效地发挥出股票市场的融资和投资功能，我国相关监管机构应该完善我国股票红利分配制度并改进我国股票市场的信息披露机制。同时，我国投资者也需加强自身对信息的认识能力，避免出现由于心理偏差而造成的投资损失。
[Abstract]:Whether the stock return rate can reflect the market information and its degree of reflection has always been concerned by the theoretical and practical circles. In theory, Fama(1973 is the earliest. In 1992, the theory of market efficiency and the three-factor model are proposed to analyze the degree of reaction of stock returns to market information and the influencing factors. In recent years, with the rise of behavioral finance, investors analyze the stock yield through fundamental and technical information to make a profit. The behavioral finance scholars, represented by de Bondtt (1985), analyze whether there is overreaction and insufficient reaction to market information from the point of view of investor psychology. Thus, the heterogeneity of stock returns to market information is demonstrated. Based on the theory of Fama and de Bondt reaction characteristics of stock returns. In this paper, we use Bayesian statistical model to empirically analyze the reaction of stock returns to open market information. Firstly, we divide the open market information into three categories, that is, enterprise financial information. Stock market information and macroeconomic information, and through AIC and Bayesian factor model selection criteria to select more significant variables to regression. Empirical results show that compared with corporate financial information. Stock market information and macroeconomic information have a greater impact on stock returns, including stock turnover rate, market risk coefficient 尾. The effect of reserve ratio and money supply on stock yield is very significant, which indicates that investors in our country are making investment decisions. We think this is mainly due to the short-term speculation of Chinese stock investors, the dividend system of stock market, and the imperfect disclosure of financial information. As to whether there is overreaction and insufficient reaction of stock investors in China, scholars used to construct "winner combination" and "loser combination". In this paper, we use Bayesian quantile regression to empirically analyze whether there is overreaction and underreaction to market information. We use the Markov chain Monte Carlo method to simulate the MCMCs and sample them by Gibbs. The results show that for those stocks whose returns have increased significantly. Investors tend to overestimate the possibility of future stock growth and exaggerate the size of the firm, turnover rate and money supply and other good news, which makes the impact of stocks on a unit of good information. Stock yields have risen more than their average, and investors have played a role in ignoring the good news for stocks that have plummeted. This makes the impact of investors on a unit of good information. The rising degree of stock yield is much lower than its average rise. So we think that investors in our country also have the overreaction and underreaction mentioned in behavioral finance. We think this phenomenon is mainly caused by investors' self-attribution bias, conservative mentality and so on. In order to make the stock market of our country play the financing and investment function of the stock market more effectively. China's relevant regulatory bodies should improve China's stock dividend distribution system and improve our stock market information disclosure mechanism. At the same time, our investors also need to strengthen their own understanding of the information ability. Avoid the loss of investment caused by psychological bias.
【学位授予单位】：上海大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.51;F224

【参考文献】