市场有效性、CAPM异象与沪深股市可预测性研究

发布时间：2018-04-22 14:17

本文选题：均值-方差理论 + FF三因素模型　；参考：《华中师范大学》2013年硕士论文

【摘要】：从有效市场假说出发,Markowitz在十九世纪五十年代建立了资产组合理论,奠定了现代金融理论的基础。随后,Sharpe(1964、Lintner(1965)以及Mossin(1966)分别独立发展了资本资产定价模型(CAPM)。他们认为,“资产收益率的差异来源于不同资产所承担的风险的差异,资产的收益越高就意味着其包含的风险越大”。CAPM有着极为严格的假设条件,因此在上世纪八十年代,其受到越来越多的实证研究的挑战,包括规模效应、长期反转效应、价值溢价效应和异质风险效应等“异象”,使得传统的CAPM理论失去了对市场收益的解释力。对此,Fama, French(1990,1992)首次将规模风险和价值风险作为风险因子,建立了著名的三因素模型。其以市场风险为系统性风险,而规模风险和价值风险为单个证券或投资组合的特征风险,在实证领域取得了较好的效果。本文参照Fama-French的三因素模型,在分组指标选择方面加以补充改进,采用股票历史贝塔值、异质风险、股票价格和公司规模四种模式进行投资组合构建,分别对上海股票市场和深圳股票市场单独进行了零贝塔形式的CAPM检验。实证结果表明：①在整个样本区间上,利用传统的历史贝塔分组方法可以得到上海和深圳股票综合指数均值-方差有效的结论。②沪深两市关于CAPM回归估计结果有所差异,相对于沪市,深市样本组合回归结果有更高的可决系数,采用零β形式的CAPM更能解释深圳证券市场股票的历史收益。③CAPM可有效地解释沪市和深市的市场风险溢价。④沪市和深市均存在规模效应和“价格效应”,小规模低价格的投资组合拥有更高的预期超额收益,深市还存在一定程度的异质波动风险溢价,预期超额收益随异质波动风险增加而减少。本文还利用引入流动性补偿因子的改进FF因素模型验证了以上结果,通过对相较于零β组合的超额收益与各溢价因子进行分析研究,发现沪市和深市的结论并不一致。沪市规模效应和“价格效应”等横截面异象完全可以由规模补偿效应、价值溢价因子和流动性补偿因子解释。而对深圳证券市场而言,异质风险溢出可以由规模补偿因子和流动性溢价因子所解释,“价格效应”主要由规模补偿效应造成,而流动性溢价效应并不是规模报酬溢出的成因。最后,本文利用该模型进行了成份股周收益率样本内可预测性研究,结果发现该模型在经济上是显著的,同时深市比沪市有更强的可预测性。以上实证结果对于中国股市中股票收益的预测、超额收益的影响因素讨论、资本资产定价等问题的研究具有一定的现实意义。
[Abstract]:Starting from the efficient market hypothesis, Markowitz established the portfolio theory in the 1850s, which laid the foundation of the modern financial theory. Then Sharpex 1964Lintnerin1965and Mossinn 1966developed the capital asset pricing model CAPMN separately. In their view, "the difference in return on assets stems from the differences in risk taken by different assets, and the higher the return on an asset means the greater the risk it contains." CAPM has extremely strict assumptions, so in the 1980s, It is challenged by more and more empirical studies, including scale effect, long-term reversal effect, value premium effect and heterogeneous risk effect, which makes the traditional CAPM theory lose its explanatory power to market returns. For this reason, fama, French 1990 / 1992), for the first time, takes the scale risk and the value risk as the risk factors, and establishes a famous three-factor model. It takes market risk as systematic risk, while scale risk and value risk are characteristic risk of single securities or portfolio, which has achieved good results in empirical field. According to the three-factor model of Fama-French, this paper supplements and improves on the selection of grouping index, and adopts four models of historical beta value, heterogeneous risk, stock price and company size to construct the portfolio. The CAPM test of zero beta form is carried out separately in Shanghai stock market and Shenzhen stock market. The empirical results show that using the traditional historical Beta grouping method, we can get the conclusion that the average variance of the composite index of Shanghai and Shenzhen stock market is effective. 2 the CAPM regression estimation results of Shanghai and Shenzhen stock markets are different. Compared with Shanghai stock market, the regression result of Shenzhen sample combination has higher determinability coefficient. Using zero 尾 CAPM can explain the historical return of Shenzhen stock market. 3CAPM can effectively explain the market risk premium of Shanghai stock market and Shenzhen stock market. 4. The scale effect and "price effect" exist in both Shanghai and Shenzhen stock markets. The small and low price portfolio has higher expected excess return, and there is a certain degree of heterogeneity volatility risk premium in Shenzhen market. The expected excess return decreases with the increase of heterogeneous volatility risk. In this paper, the improved FF factor model with liquidity compensation factor is used to verify the above results. Through the analysis of excess return and premium factors compared with zero 尾 combination, it is found that the conclusions of Shanghai Stock Exchange and Shenzhen Stock Exchange are not consistent. The cross-sectional anomalies such as scale effect and "price effect" in Shanghai stock market can be explained by scale compensation effect, value premium factor and liquidity compensation factor. For Shenzhen stock market, heterogeneous risk spillover can be explained by scale compensation factor and liquidity premium factor. "Price effect" is mainly caused by scale compensation effect, and liquidity premium effect is not the cause of scale reward spillover. Finally, this paper uses the model to study the predictability in the component stock weekly return sample. The results show that the model is significant in economy, and the Shenzhen stock market is more predictable than Shanghai stock market. The above empirical results are of practical significance to the prediction of stock returns in Chinese stock market, the discussion of influencing factors of excess returns, and the pricing of capital assets.
【学位授予单位】：华中师范大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F224;F832.51

【参考文献】