中国股市收益率偏度特性研究

发布时间：2017-12-31 15:29

本文关键词：中国股市收益率偏度特性研究　出处：《复旦大学》2014年硕士论文　论文类型：学位论文

【摘要】：在资本市场中,资产收益率分布经常表现出不对称性,即偏度性,并且关于偏度现象的研究已经有近四十年的时间了。然而,在实际应用和研究中经常假设资产收益率服从正态分布等对称性分布,忽视了偏度特性。如果偏度现象真实存在,且不容忽视,那么简单的正态分布假设极有可能造成对真实收益率和风险特征认识的误导。众所周知,收益率影响着资产定价、投资组合、风险预测和市场参与者行为等诸多方面,偏度作为资产收益率的固有属性,进而影响着资产定价,也决定着配置资产以及影响着市场参与者行为和监管等等。因此,研究资本市场收益率的偏度特性具有极大的理论意义和实践意义。文章以2006年到2013年间的中国股票市场为研究对象,基于不同分布构建的随机波动模型(SV),研究论证了股市收益率偏度现象的存在性及其主要特点。其中,八种不同概率分布分别是：正态分布(N)、偏斜正态分布(SN)、学生t分布(T)、偏斜学生t分布(ST)、广义误差分布(GED)、偏斜广义误差分布(SGED)、混合正态分布(MN)以及混合偏斜正态分布(MSN)。通过实证研究,得到了已下几点结论：(1)正态分布的拟合效果最差,收益率服从正态分布的假设并非最优选择；(2)SN和N、ST和T、SGED和GED以及MN和MSN四对八个分布中,每对分布中含偏度参数的分布至少不会比相对应不含偏度参数的分布表现得差；(3)相比N和SN分布,具备刻画厚尾能力的分布表现得更好,说明中国股票市场不仅存在偏度现象,也伴随着普遍的尖峰厚尾性；(4)综合比较,混合偏斜正态分布(MSN)拟合效果最好,它不仅能够抓住股市收益率的偏度性和厚尾性,而且还可以分别刻画左、右尾部的“非常态”收益率情况。最后,我们以资本资产定价模型(CAPM)为一个例子,展示了偏度的应用价值意义。我们在CAPM模型中引入最好的MSN分布,以研究偏度对资本资产定价的影响,实证结果认为：基于正态分布假设的经典CAPM模型限制了α的真实值,如果用于回归具有负偏形态的各行业指数收益率,得到的结论其实是高估了行业真实收益率、低估了风险。总之,如果没能够充分考虑收益率的偏度和厚尾性,极有可能对资产的真实收益率和风险情况产生错误的认识。在偏度的市场中,选择相应的偏斜的分布研究金融问题,具有普遍的理论意义和现实意义。比如在资产定价、投资组合构建、期货交易保证金比例设定、金融风险预测以及监管规则制定等等方面都大有用处。
[Abstract]:In the capital market, the distribution of return on assets often shows asymmetry, that is, skewness, and the research on skewness has been for nearly forty years. In the practical application and research, it is often assumed that the symmetrical distribution of the return rate of assets from normal distribution, ignoring the skewness characteristics, if the skewness exists, and can not be ignored. That simple assumption of normal distribution is likely to lead to misleading perceptions of the true rate of return and the characteristics of risk. As we all know, the rate of return affects asset pricing and portfolio. Risk forecasting and market participants' behavior, as the inherent attribute of asset return rate, affect asset pricing. It also determines the allocation of assets and influences the behavior and regulation of market participants and so on. It is of great theoretical and practical significance to study the skewness of capital market returns. This paper takes the Chinese stock market from 2006 to 2013 as the research object. Based on the stochastic volatility model constructed with different distributions, this paper studies the existence and main characteristics of the skewness of stock market returns. Among them, eight different probability distributions are normal distribution. Skew normal distribution, student t distribution, skew student t distribution, generalized error distribution and skew generalized error distribution. Mixed normal distribution (MN) and mixed skew normal distribution (MSN). Through empirical research, we have got the following conclusions: 1) the fitting effect of normal distribution is the worst. The assumption that the rate of return is normal distribution is not the best choice; SGED and GED and MN and MSN in four pairs of eight distributions. The distribution of biasing parameters in each pair of distributions is at least not worse than that of the corresponding distribution without bias parameters. Compared with the distribution of N and SN, the distribution with the ability to depict the thick tail is better, which indicates that the stock market in China has not only the phenomenon of bias, but also accompanied by the widespread spike and thick tail. Comprehensive comparison shows that the mixed skew normal distribution has the best fitting effect. It can not only grasp the skewness and thick tail of the stock market yield, but also depict the left. The "abnormal" yield at the right end. Finally, we take the capital asset pricing model (CAPMM) as an example. We introduce the best MSN distribution in CAPM model to study the influence of bias on capital asset pricing. The empirical results show that the classical CAPM model based on the normal distribution hypothesis limits the true value of 伪, if it is used to return the returns of industries with negative bias. The conclusion is that it overestimates the real yield of the industry and underestimates the risk. In a word, if we can not fully consider the skewness and thick tail of the yield. It is very likely to have a wrong understanding of the real return rate and risk of assets. In the biased market, choose the corresponding skew distribution to study the financial problems. It has universal theoretical and practical significance, such as asset pricing, portfolio construction, margin ratio setting of futures trading, financial risk forecasting and regulatory regulation.
【学位授予单位】：复旦大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F832.51

【相似文献】