基于投资者情绪的行为资产定价研究

发布时间：2018-06-02 14:50

本文选题：投资者情绪 + 行为资产定价模型　；参考：《华南理工大学》2012年博士论文

【摘要】：现代金融理论产生后得到了长足的发展，然而，自70年代末以来涌现出大量投资者异常行为及金融市场异象，从而使得现代金融理论的完备体系面临了金融事实的挑战。自此，行为金融学产生并发展开来。对于行为金融领域中的重要问题，即资产定价模型而言，现有文献主要从噪音交易者和投资者认知偏差两个角度进行了研究。然而，基于投资者情绪的研究与基于这两个研究角度的分析相比，具有更大的优势，既能较容易实证检验相关结论，也能得到心理行为实验与神经医学实验的证据支持。因此，从投资者情绪的角度进行行为资产定价模型的研究或许能够得到更多的有益结论。本文即基于投资者情绪理论研究行为资产定价问题，继而深入分析所构建的理论模型，文章主要研究内容有三：一、针对目前投资者情绪指数构建过程中面临的一些不足，本文提出了一套选取投资者情绪代理变量的优化程序。本文提出的优化程序对原始代理变量的选取做了改进，同时对不适合表征情绪的代理变量增加了剔除操作，，从而使得构建的投资者情绪指数更加符合理论逻辑。随后，本文应用我国股市实际数据，以EGARCH模型验证了该优化情绪指数的有效性，从而为本文理论模型的研究奠定了坚实的基础。二、基于BSV模型与DSSW模型，本文相继构建了风险资产的情绪认知价格模型和情绪均衡价格模型，并数理推导了其价格解析表达式。首先，针对市场上仅存在风险资产、以及市场情绪投资者同质即均为单向情绪投资者的假设条件下，本文改进了BSV模型的不足之处，建立了单向情绪资产认知价格模型，并通过数理推导求解得到了其价格表达式。随后将模型拓展至市场情绪投资者异质即可分为正向情绪投资者与负向情绪投资者两类时的情况，构建并求解了双向情绪资产认知价格模型。其次，本文考虑了当市场上同时存在风险资产和无风险资产的情况下，投资者在受到情绪影响的条件下如何配置其资产并最终决定风险资产的价格，在改进了DSSW模型的基础上分别构建了单向情绪资产均衡价格及双向情绪资产均衡价格模型，并通过推导求得了两模型的解析表达式。三、针对文中两类理论模型求解得出的解析表达式，本文分析了实际投资策略、资产超额收益、资产情绪泡沫及负的期望收益等问题，并得到了丰富的有益结论。针对两类情绪认知价格模型，特别是单向情绪认知价格模型，设计了一套投资策略，利用本文提出的投资者情绪指数构造优化程序，构建了一类新的市场整体情绪指数，并以上证综指为对象进行了实际数据的验证，发现该投资策略相较于基准指数能够获取较高的超额收益，从而验证了本文所提出投资策略的有效性。随后参数灵敏度分析证明了初始参数的设置并不影响投资策略的盈利率，最终又验证了这一投资策略盈利的稳健性。对于双向情绪认知价格模型，本文给出了数值模拟，并以参数分析验证了众多行为金融实验的结论，即正向情绪投资者与负向情绪投资者权衡博弈时，正向情绪投资者数量的增加将引致市场平均情绪水平高涨，并最终导致资产价格升高。针对两类情绪均衡价格模型，本文分别分析了情绪对风险资产超额收益、泡沫及期望收益的影响。在单向情绪资产均衡价格模型方面，研究结论表明：情绪高涨时资产有正的期望收益，情绪低落时资产有负的期望收益；然而，这一结论对超额收益和泡沫而言并不成立，即当情绪低落时，风险资产确实会有正的超额收益及负的情绪泡沫，但当情绪高涨并大于某一临界值时，即情绪剧烈高涨时，资产具有负的超额收益及正的情绪泡沫，当情绪高涨小于该临界值时，即情绪温和高涨时，资产具有正的超额收益及负的情绪泡沫。这一数理推导观点是对现有研究结论的细化。在双向情绪资产均衡价格模型方面，研究结论表明：负向情绪投资者的悲观气氛强于正向投资者的乐观气氛时，风险资产有超额收益；正向情绪投资者的乐观气氛强于负向情绪投资者的悲观气氛但强度不是很大即小于某一临界值时，风险资产有超额收益；正向情绪投资者的乐观气氛远远强于负向情绪投资者的悲观气氛即大于该临界值时，风险资产没有超额收益。对泡沫的分析与对超额收益的分析均得到了相同的情绪临界值，这一结论的性质与前面的单向情绪资产均衡价格模型分析一致。最后分析了双向情绪水平的变化与投资者数量的变化分别对资产期望收益的影响，结果发现当两类情绪发生同向变化且低落时资产期望收益为负，这一情况意味着市场整体情绪水平的低落，与单向情绪资产价格模型的结论是一致的；当两类情绪发生异向变化时资产期望收益的正负取决于价格模型中其它参数的具体取值情况。而乐观情绪投资者数量的增加并不一定使得风险资产价格必然地上升，该类投资者的过度涌现可能造成资产价格的风险抑制效应增强，从而可能使其具有负的期望收益，这与实际市场中的股票价格行为一致。
[Abstract]:The modern financial theory has developed greatly. However, since the end of the 70s, a large number of abnormal behavior and financial market anomalies have emerged, which makes the complete system of modern financial theory face the challenge of financial facts. In the case of asset pricing model, the existing literature is mainly studied from two perspectives, the noise trader and the investor's cognitive bias. However, the research based on investor sentiment has a greater advantage than the analysis based on these two perspectives. Therefore, the study of behavioral asset pricing model from the perspective of investor sentiment may get more useful conclusions. This paper is based on investor sentiment theory to study behavior asset pricing, and then analyzes the theoretical model constructed in depth. The main contents of this paper are three:
First, in view of the deficiencies in the construction of investor sentiment index, this paper proposes a set of optimization procedures to select investor sentiment agent variables.
The optimization program proposed in this paper improves the selection of the original proxy variables, and increases the elimination operation for the agent variables that are not suitable for the representation of emotions, thus making the investor sentiment index more consistent with the theoretical logic. Then, this paper uses the actual data of the stock market in China to verify the optimization of the emotional index with the EGARCH model. Efficiency, thus laying a solid foundation for the theoretical model of this paper.
Two, based on the BSV model and the DSSW model, this paper constructs the emotional cognitive price model and the emotional equilibrium price model of the risk assets, and deduces the analytic expression of its price.
First, under the assumption that there is only a risk asset in the market, and the market sentiment investor is homogeneity as a one-way emotional investor, this paper improves the deficiency of the BSV model, establishes a one-way emotional asset cognitive price model, and obtains its price expression by mathematical derivation. Then the model is extended to the market. The heterogeneity of emotional investors can be divided into two groups of positive emotional investors and negative emotional investors, and the cognitive price model of two-way emotional assets is constructed and solved.
Secondly, we consider how to configure their assets and ultimately determine the price of risk assets under the condition of emotional impact, when the market exists both risk assets and riskless assets. On the basis of the improved DSSW model, the equilibrium price of one-way mood assets and the equilibrium price of two-way emotional assets are constructed respectively. The analytic expression of the two model is obtained through deduction.
Three, in view of the analytical expressions of the two types of theoretical models, this paper analyzes the actual investment strategy, the excess return on assets, the asset sentiment bubble and the negative expected return, and has obtained a rich and useful conclusion.
According to the two kind of emotional cognitive price model, especially one way emotion cognitive price model, a set of investment strategy is designed. By using the investor sentiment index structure optimization program, a new kind of market overall emotional index is constructed, and the above index is verified by the actual data, and the investment strategy is found to be compared. The Yu Jizhun index can obtain high excess returns, thus verifying the effectiveness of the investment strategy proposed in this paper. Then the parameter sensitivity analysis proves that the setting of initial parameters does not affect the earnings rate of investment strategy, and finally verifies the robustness of the investment strategy. The numerical simulation is used to verify the conclusion of many behavioral financial experiments, that is, when the positive emotional investors and negative emotional investors weigh the game, the increase of positive emotional investors will lead to higher market average emotional level, and ultimately lead to higher asset prices.
According to the two kind of emotional equilibrium price model, this paper analyzes the effect of emotion on excess return, bubble and expected return. In the aspect of equilibrium price model of one way emotional asset, the conclusion shows that the asset has positive expected return and negative expected return when the mood is low; however, this conclusion is a conclusion. It is not true for excess returns and bubbles, that is, when the mood is low, the risk asset does have positive excess returns and negative emotional bubbles, but when the mood is higher than a critical value, that is, when the emotion rises sharply, the assets have negative excess returns and positive emotional bubbles, when the sentiment is higher than the critical value, that is, emotion. At the time of mild inflation, the assets have positive excess returns and negative emotional bubbles. This mathematical derivation view is a refinement of the existing research conclusions. In the aspect of the equilibrium price model of two-way emotional assets, the conclusion shows that the negative mood of negative emotional investors is stronger than positive investors, and the risk assets have excess returns. Positive mood investor's optimistic atmosphere is stronger than negative mood investor's pessimistic atmosphere, but the risk asset has excess return when the intensity is not large or less than a certain critical value, and the positive mood of positive emotional investor is far stronger than negative mood investor's pessimistic atmosphere is larger than the critical value, and the risk asset has no excess return. The analysis of bubbles and the analysis of excess returns all get the same emotional critical value. The nature of this conclusion is consistent with the analysis of the equilibrium price model of the one-way emotional asset. Finally, we analyze the effect of the change of two-way emotion level and the change of investor quantity on the expected return of the investment period respectively. The results are found to be the two kinds of emotions. In the same direction, the expected return of assets is negative, which means that the overall sentiment level of the market is low and the conclusion of the one-way emotional asset price model is consistent. The positive and negative value of the expected income of the assets depends on the specific value of the other parameters in the price model when the two kinds of emotions vary. The increase in the number of emotional investors does not necessarily increase the price of the risk assets. The excessive emergence of this kind of investor may cause the increase of the risk inhibition effect of the asset price, which may lead to the negative expected return, which is consistent with the stock price behavior in the real market.
【学位授予单位】：华南理工大学
【学位级别】：博士
【学位授予年份】：2012
【分类号】：F832.51;F224

【引证文献】