股指收益波动的非线性与异方差性

发布时间：2018-05-27 00:11

本文选题：股票波动 + GARCH模型　；参考：《广西师范大学》2013年硕士论文

【摘要】：Campbell和Mackinlay(1997年)提出,许多经济行为都不是线性的,他们认为信息的不对称、信息的缺乏和市场的不完善性和市场微观结构的特征导致了对新信息的滞后反应,从而导致了在股票价格变化中的非线性,同时Hsieh(2003年)发现股票收益呈现出了非线性依赖以及条件异方差可以解释其非线性。 2008年中国的股市的崩盘吸引了很多学者对股票市场进行深入分析,通过分析股票收益的风险性和不确定性的敏感程度来达到对股票市场的一个预期,帮助投资者更加理性的进行投资,现代投资理论认为投资者应该基于风险程度和不确定性做出投资决策。如何对股票收益波动率进行准确的描述与预测?近年来伴随着金融风暴的再次冲击而重新上升为金融学领域探讨的热点问题。投资者通过掌握股票收益波动率的特征及趋势来达到对股票市场中风险的测度、规避和管理帮助更好的进行投资。因此,掌握股票收益波动率的特征及趋势具有极其重要的理论和实际意义。本文研究了中国股票收益的非对称性和股票收益波动的异方差性。本文进一步的分析了股票收益、条件方差以及标准残差之问的关系。本文选取了2002年到2012年的沪深300指数、上证指数和深成指数日开盘价格和收盘价格作为研究数据,选用了Liung-Box统计指标检验了股票收益的自相关性,同时通过GARCH模型和TAR-GARCH(1-1)模型分别检验了股票收益的异方差性和股票收益的非对称性。本文的研究结果阐述了现阶段我国股票收益存在着非线性、异方差效应以及非对称性。同时也发现股票收益与条件波动之间没有相关性,但是股票收益与标准残差之间存在着显著地正相关性。这些发现为现代投资理论引入了一个关键的元素,即投资者应根据其预期波动来调整他们的投资决策,然而事实上,他们更趋向在承担非预期波动同时收到额外的风险报酬。本文引入了稳健性检验来对条件方差进行检验,同时研究了利好消息和利空消息的非对称影响。到目前为止,国内利用AR模型、GARCH模型、EGARCH模型来解决实际问题的文献有很多,但在运用到我们股票市场的模型却为数不多,特别是用TAR-GARCH模型,并且在选取数据方面,绝大多数文献均是选取上证指数、深成指数或者是单独的创业板指数、中小板指数,但是这些并不能反映整个市场的走势,而本文选取的是沪深300指数、上证指数和深成指数作为数据来源,本文试图通过衡量其非线性进而测算异方差对股票市场波动的影响。本文同时检验了波动的非对称性。实证结果指出了在预测中国股票市场收益的条件波动中GARCH (1,1)模型最为合理。本文还发现其波动性明显依赖于历史误差项,即前期收益和波动的非预期上升和下降对波动性影响显著。这是因为历史时期的股票收益和波动的非预期上升和下降共同影响了投资者行为,从而影响了他们的投资决策。本文同时发现利好和利空消息对股票市场波动的影响具有非对称性。股票收益和非预期波动之间的显著正相关性表明投资者在股票收益非预期的上升和下降阶段得到风险补偿。预期波动理论建议投资者根据他们的预期波动调整他们的投资组合,本文研究结果表明股票收益和预期波动之间存在正相关关系,但非预期波动对股票收益的影响更大。我国股票市场上“跟风"现象严重,市场投资者的投资和风险理念还很不成熟。当市场运行并不具备有效性,信息传递存在极大障碍和代价时,政府的市场干预便具备合理的经济学基础。因此,在我国股市发展的初期阶段,就必须特别强调政府对股市政策的建立与操作,避免出现股价过度波动和资源无效分配以及可能产生的金融恐慌与危机。股市监管的重点应该放在保护投资者利益,健全完善信息批露制度,保证信息公平获取权,惩处内幕信息内幕交易虚假陈述等行为,促使信息在股市迅速准确对称流动,促进股市稳健的发育与完善,同时投资者也应加强自身学习,提高自身素质,尽量做到合理有效投资。
[Abstract]:Campbell and Mackay ( 1997 ) suggest that many economic behaviors are not linear , they believe that the asymmetry of information , the lack of information and the imperfection of the market and the characteristics of the market microstructure lead to the hysteresis of the new information , which leads to the nonlinearity in the stock price change , and Hsieh ( 2003 ) finds that the stock return presents nonlinear dependence and conditional heteroscedasticity can explain its non - linearity .

In 2008 , China ' s stock market crash attracted many scholars to analyze the stock market . By analyzing the sensitivity of the risk and uncertainty of stock return , investors should make investment decision based on the degree of risk and uncertainty . How to accurately describe and forecast the fluctuation rate of stock return ?

In this paper , we study the relationship between the non - symmetry of stock return and the volatility of stock return . The paper further analyzes the relationship between stock return , conditional variance and standard residual . The paper selects Liung - Box statistic index to test the self - correlation of stock return .

So far , there are a lot of literatures to solve the real problems by using the AR model , the ARCH model and the EGARCH model in China , but the model of the stock market is limited , especially the TAR - ARCH model , and most of the literatures are used as the data source . The paper also examines the asymmetry of the volatility . The empirical results point out that the ARCH ( 1,1 ) model is the most reasonable in predicting the conditional volatility of the stock market returns .

This paper also finds that the volatility of stock returns and volatility is significantly influenced by historical error terms , that is , the unexpected rise and fall of stock returns and volatility in the previous period have a significant impact on volatility . This is because the positive correlation between stock return and unexpected volatility has a positive correlation between stock return and unexpected volatility . The results of this paper suggest that investors have positive correlation between stock returns and expected fluctuations , but the impact of non - expected volatility on stock returns is greater .

In the early stage of the development of stock market , the government must emphasize the government ' s establishment and operation of stock market policy , avoid excessive fluctuation of stock price and ineffective distribution of resources and the possible financial panic and crisis .
【学位授予单位】：广西师范大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.51;F224

【参考文献】