基于投资者情绪效应的均值—方差关系模型研究

发布时间：2018-01-05 20:36

本文关键词：基于投资者情绪效应的均值—方差关系模型研究　出处：《广州大学》2016年博士论文　论文类型：学位论文

【摘要】：关于金融市场收益、风险影响因素的研究一直是学者们所追逐的热点之一.近年来,行为金融学的蓬勃发展为众多学者们提供了解决金融问题的新思路.行为金融学强调投资者心理、情绪因素在投资决策与市场定价中的作用,由此它成功地应用心理学中的成果解释了金融市场的许多异象并解决了一些难点问题.但是金融问题的研究不能局限在定性描述与历史观察中,需要缜密的逻辑推断和统计模型的论证选取来定量描述投资者心理是如何影响市场的.因为Markowitz用数学期望和标准差度量收益与风险,所以在金融领域中通常将市场收益风险关系也称作“均值方差”关系.鉴于金融市场收益风险关系普遍存在时变性的事实,本文立足于“投资者情绪”的视角,利用统计方法构建模型(以均值方差模型为主)并对其进行探讨.本文主要内容包括如下三个方面:第一,从决定投资者情绪的心理偏差、偏好等行为因子出发,我们构建了分析投资者情绪对投资绩效影响的结构方程模型,这些模型包括含交互项的结构方程模型、纳入协变量的结构方程模型及多组结构方程模型.我们采用实际调查所搜集的数据,通过实证检验发现:投资者的具体情绪包括它们的交互项会对投资绩效产生显著的影响;情绪的表现程度以及对绩效的影响程度会由个体特征(如性别、教育程度等)不同而不同;机构投资者与个体投资者所表现的效应也会有所差异.第二,依据投资者心理会引起市场中收益风险关系的时变性,我们建立了一类带有潜变量的变系数ARCH-M模型.考虑到潜变量的不可直接观测性,我们采用校正似然方法对参数进行了估计,并且证明在一些正则条件下,估计是相合且渐近正态的.我们通过建立一个检验统计量完成了潜变量效应的检验.同时数值模拟结果表明我们的估计和检验具有良好的性质.此外,我们还应用所构造的模型实证检验了在真实市场中收益风险关系是受到投资者情绪影响的.第三,通过构建带有GARCH效应的门限模型,我们实证研究了投资者情绪对市场收益风险关系的影响.结果表明,市场收益风险关系受投资者情绪的影响是存在转变机制的.当市场情绪落入低迷期,两者关系为显著负相关;当投资者情绪进入到高涨期时,两者关系是显著正相关;当情绪处在复苏、稳定期时,两者关系并不显著,而且当期情绪的这种作用还会影响到下期的市场.同时我们发现股市政策性的变化会引起情绪波动.
[Abstract]:About the financial market income, the risk influence factor research has been one of the hot spots pursued by the scholars. In recent years. The vigorous development of behavioral finance provides many scholars with a new way to solve financial problems. Behavioral finance emphasizes the role of investor psychology and emotional factors in investment decision-making and market pricing. Therefore, it has successfully applied the achievements in psychology to explain many anomalies of financial market and solve some difficult problems, but the study of financial problems cannot be confined to qualitative description and historical observation. Careful logical inference and statistical model selection are needed to quantitatively describe how investor psychology affects the market because Markowitz measures returns and risks with mathematical expectations and standard deviations. Therefore, in the financial field, the relationship between market returns and risk is usually referred to as the "mean variance" relationship. This paper is based on the perspective of "investor sentiment", using statistical method to build a model (mainly mean variance model) and discuss it. The main content of this paper includes the following three aspects: first. Based on the behavioral factors, such as psychological bias, preference and so on, we construct structural equation models to analyze the impact of investor sentiment on investment performance. These models include structural equation models with interaction terms. The structural equation model and multiple sets of structural equation models are included in the covariates. We use the data collected from the actual investigation. The empirical results show that investors' specific emotions, including their interactions, have a significant impact on investment performance; The degree of expression of emotion and the degree of influence on performance will be different from individual characteristics (such as gender, education, etc.); The effects of institutional investors and individual investors will also be different. Secondly, according to investor psychology, it will cause the time-varying of the relationship between income and risk in the market. We establish a class of variable coefficient ARCH-M model with latent variables. Considering the non-direct observability of latent variables, we estimate the parameters by means of correction likelihood method. And it is proved that under some regular conditions. The estimation is consistent and asymptotically normal. By establishing a test statistic, we have completed the test of the latent variable effect. The numerical simulation results show that our estimation and test have good properties. We also use the constructed model empirical test in the real market income risk relationship is affected by investor sentiment. Third, by building a threshold model with GARCH effect. We empirically study the influence of investor sentiment on the relationship between market returns and risk. The relationship between market returns and risk is influenced by investor sentiment. When market sentiment falls into a downturn, the relationship between them is significantly negative. When investor sentiment enters the period of upsurge, the relationship between the two is significantly positive correlation; When the emotion is in the recovery and stable period, the relationship between the two is not significant, and the effect of the current emotion will also affect the market in the next period. At the same time, we find that the change of the stock market policy will cause the emotional fluctuation.
【学位授予单位】：广州大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：F224;F830

【相似文献】