股票收益率及系统性风险的估计研究

发布时间：2018-01-04 18:29

本文关键词：股票收益率及系统性风险的估计研究　出处：《重庆大学》2016年硕士论文　论文类型：学位论文

【摘要】：文章主要讨论的问题是资本市场中股票收益率的波动性建模以及系统风险系数的动态估计。研究的两个重点是股票收益率和系统风险系数,收益率可以很好的衡量某个资本单日在前一日基础上的收益情况,同时可以很好的衡量股票的波动性。系统风险β系数衡量的是单个资本受到的市场的影响,反映证券的收益率水平对市场平均收益水平变化的敏感度,是衡量证券承担系统风险水平的指标。投资组合理论中的系统风险系数在理论研究以及投资实践中都具有非常重要的作用,对它的研究将有效的为资产定价以及风险管理提供决策依据。文章的研究主要运用的工具是时间序列分析方法中GARCH族模型和多元GARCH模型。金融时间序列一般情况下呈现出阶段性的相对平稳和阶段性的剧烈波动,因此采用波动性建模的ARCH模型族和GARCH模型族进行估计。另外,在研究系统风险系数时需要估计单只股票与市场指数间的动态条件相关系数序列,需要运用多元GARCH模型,文章选取了DCC-MVGARCH两步建模法实现了对动态条件相关系数序列的估计。文章选取了上证50指数及其成分股的日线交易数据作为样本,在完成对收益率数据的非正态性检验、稳定性检验和ARCH效应检验的基础上对数据进行建模。在比较了模型的AIC、SIC、R平方和残差平方和等指标后确定了最优的模型为GARCH模型族中适用于非对称建模的EGARCH(2,2)模型,实现了对上证50指数收益率序列的良好拟合。最后文章收益率序列建模的基础上讨论了时变的系统性风险β系数。根据DCC-MVGARCH模型获得成分股收益率与指数股收益率的动态相关系数序列以及各自的条件方差序列,带入模型βi=cov(ri,rn)/σm2从而得到系统风险系数序列。研究结果表明大多数成分股收益率的β系数都在0.5-1.5之间波动,且总体而言,系数在2015年二季度时波动变小,这段时间各个股票走势与大盘比较接近。所有成分股中,贵州茅台的β系数数值较小且较为稳定,是上证50指数板块中风险最小的优质蓝筹股,而招商证券的β系数较大,可能存在相当高的风险。
[Abstract]:This paper mainly discusses the volatility model of stock return in capital market and the dynamic estimation of system risk coefficient. The two emphases of the study are stock return and systematic risk coefficient. The yield can well measure the return on a single capital day on the basis of the previous 1st, and can also measure the volatility of the stock. The systematic risk 尾 coefficient measures the impact of the market on a single capital. The sensitivity of the yield level of the securities to the change of the average return level of the market. The systematic risk coefficient in portfolio theory plays an important role in both theoretical research and investment practice. The main tools of this paper are GARCH family model and multivariate GARCH model in time series analysis. The time series generally show a relatively stable phase and violent fluctuations of the phase. Therefore, the volatility modeling ARCH model family and the GARCH model family are used to estimate the dynamic conditional correlation coefficient series between a single stock and the market index. Multivariate GARCH model is needed. This paper selects DCC-MVGARCH two-step modeling method to realize the estimation of dynamic conditional correlation coefficient series, and selects the daily trading data of Shanghai Stock Exchange 50 Index and its constituent stocks as samples. The data are modeled on the basis of non-normal test, stability test and ARCH effect test, and the AIC-SIC of the model is compared. After R squared sum and residual squared sum, the optimal model is EGARCH2 / 2) model which is suitable for asymmetric modeling in GARCH model family. The good fitting of the yield series of Shanghai 50 index is realized. Finally, the time-varying systematic risk 尾 coefficient is discussed on the basis of the model of return sequence. The components are obtained according to the DCC-MVGARCH model. The dynamic correlation coefficient series of stock yield and index stock return and their conditional variance series. A series of systematic risk coefficients are obtained by introducing the model 尾 -covrigne / 蟽 m ~ 2. The results show that the 尾 coefficients of most component stock returns fluctuate between 0.5-1.5. In general, the volatility of the coefficient in the second quarter of 2015 became smaller, during this period the trend of each stock is close to the market. Among all the constituent stocks, the 尾 coefficient of Moutai in Guizhou is smaller and more stable. It is the least risky high-quality blue chip in the Shanghai 50 index, while the 尾 coefficient of China Merchants Securities is large, which may have a high risk.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：F830.91;F224

【参考文献】