高频数据下投资者情绪与股票收益的研究

发布时间：2018-05-16 13:27

本文选题：条件资产定价模型 + 非参数估计方法　；参考：《浙江工商大学》2013年硕士论文

【摘要】：资本资产定价模型认为,市场风险是决定股票价格的唯一因素,市场风险涵盖了资本市场存在的所有风险。然而随着越来越多金融异象的出现,人们开始对资本资产定价模型产生怀疑。Fama和French通过对美国股市数据的实证研究提出了经典的三因子定价模型,认为市场因子、规模因子、价值因子共同影响了资产的价格。然而三因子模型仍然不能解释股票市场存在的动量效应。有效市场假说的支持者研究发现,资产定价模型无法解释越来越多的金融异象并不是因为市场无效,而是因为模型忽略了风险的动态变化趋势,于是资产定价模型开始逐渐向时变的条件资产定价模型发展。本文采用非参数估计方法,利用基于数据驱动的方法寻找最优估计窗宽,对中国A股市场2000年-2011年日度数据进行了FF三因子模型的实证研究,通过将非参数估计结果与经典的滚动窗口估计结果对比,发现非参数方法明显优于滚动窗口估计,通过对时点定价误差和长期定价误差的检验,发现非参数估计方法更适合中国股票市场；而后我们对三因子模型进行了股票市场上下行定价因子的检验,发现市场因子和规模因子通过了检验,而价值因子并没有通过检验。这说明在中国股票市场,只有市场因子和规模因子是定价因子,相应的系统风险影响资产组合的价格;而价值因子是特别风险,市场不会一直对其进行补偿,价值因子在中国并不是定价因子。基于上述的实证结果,本文创造性地采用了对中国股票市场价格有很大影响力的投资者情绪指标替代没有通过定价因子检验的价值因子,并对2000年-2011年的中国A股日度数据进行了实证研究。实证表明,改进的三因子模型虽然在实证效果方面弱于FF三因子模型,但是改进的三因子模型通过了定价因子的检验,表明投资者情绪因子在中国股票市场是定价因子,相应的风险属于系统风险,影响资产组合的价格。
[Abstract]:According to the capital asset pricing model, market risk is the only factor that determines the stock price, and the market risk covers all the risks existing in the capital market. However, with the emergence of more and more financial anomalies, people began to doubt the capital asset pricing model. Fama and French put forward the classic three-factor pricing model through empirical research on American stock market data. Value factors affect the price of assets together. However, the three-factor model still can not explain the momentum effect in stock market. Research by proponents of the efficient Market hypothesis has found that asset pricing models fail to explain more and more financial anomalies not because markets are ineffective, but because the models ignore the dynamic trends of risk. So the asset pricing model began to develop to the time-varying conditional asset pricing model. In this paper, we use non-parametric estimation method and data-driven method to find the optimal estimation window width, and carry out an empirical study of the FF three-factor model for the daily data of China A-share market from 2000 to 2011. By comparing the non-parametric estimation results with the classical rolling window estimation results, it is found that the non-parametric method is obviously superior to the rolling window estimation, and the time point pricing error and the long-term pricing error are tested. It is found that the non-parametric estimation method is more suitable for the Chinese stock market, and then we test the three-factor model for the up-down pricing factor of the stock market, and find that the market factor and the scale factor have passed the test, but the value factor has not passed the test. This shows that in the Chinese stock market, only market factors and scale factors are pricing factors, and the corresponding systemic risk affects the price of the portfolio; and the value factor is a special risk, and the market will not always compensate for it. The value factor is not a pricing factor in China. Based on the above empirical results, this paper creatively uses the investor sentiment index which has great influence on the Chinese stock market price to replace the value factor which has not passed the pricing factor test. The daily data of Chinese A shares from 2000 to 2011 are studied empirically. The empirical results show that the improved three-factor model is weaker than FF three-factor model in empirical effect, but the improved three-factor model passes the test of pricing factor, which indicates that investor sentiment factor is a pricing factor in Chinese stock market. The corresponding risk belongs to the system risk, which affects the price of the portfolio.
【学位授予单位】：浙江工商大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F224;F832.51

【参考文献】