基于贝叶斯Markov转换模型的股市收益与通胀动态关系研究

发布时间：2018-04-20 02:13

本文选题：股票收益率 + 通货膨胀率　；参考：《湖南大学》2014年硕士论文

【摘要】：通货膨胀作为制定相关宏观经济政策的重要参考指标，其波动特征和未来走势对经济形势的发展有重要的影响。股票市场作为金融市场的代表，其周期波动会对社会货币总供给与货币总需求产生结构性的影响，从而影响通胀率的波动。因此，研究股票市场与通货膨胀率之间的关系，有助于我国宏观经济政策的正确制定，并对判断市场经济政策的效应具有重要意义。通货膨胀与股票实际收益率之间关系不是纯粹的只有一种形式，其可能存在四种相关关系：正相关、负相关、不确定与不相关，实际的市场经济中股市收益与通货膨胀率之间一般呈现出非线性、非对称性的关系。Markov机制转换VAR模型（MSVAR）在研究非线性动态关系中有很广泛的应用，因为Markov机制转换能很好地刻画出时间序列的非线性机制转换的过程，然而，MSVAR模型却存在对数据依赖程度高、估计参数过多、拟合程度低等问题，贝叶斯统计推断方法以及MCMC算法的运用为解决以上这些问题提供了良好的解决方案。本文首先针对股市收益与通胀波动关系分析过程中随机参数条件下的建模问题，构建贝叶斯Markov转换VAR模型。然后，通过分析模型的统计结构，设置参数的先验概率分布，利用贝叶斯统计方法，推断模型参数的后验分布，设计相应的两次Gibbs抽样算法对模型参数进行贝叶斯估计。最后，利用贝叶斯Markov转换VAR模型研究股票回报率，通胀率波动成分和通胀率趋势成分三者间的动态关系，研究结果表明：贝叶斯Markov转换VAR模型更准确地刻画出三变量之间波动关系的非线性动态特征。在机制转移的具体过程中，股票收益率在“市场紧缩区制”分别与持久性通胀率，，暂时性通胀率呈正相关，在“市场扩张区制”短期内与通胀率呈弱正相关，长期内与通胀率呈负相关，表明了费雪效应和代理效应分别在市场的不同机制中体现出来。说明了股票收益率与通胀率的波动关系并不总是遵循费雪效应，可能存在多种关系。
[Abstract]:Inflation is an important reference index for the formulation of relevant macroeconomic policies, and its fluctuation characteristics and future trends have an important impact on the development of the economic situation. As the representative of financial market, the periodic fluctuation of stock market will have a structural impact on the total supply of money and the aggregate demand of money, thus affecting the fluctuation of inflation rate. Therefore, the study of the relationship between stock market and inflation rate is helpful to the correct formulation of macroeconomic policy in China, and it is of great significance to judge the effect of market economy policy. The relationship between inflation and real return of stocks is not purely one form, it may have four kinds of correlation: positive correlation, negative correlation, uncertainty and non-correlation, In the actual market economy, the relationship between stock market income and inflation rate is generally nonlinear and asymmetric. Markov mechanism transformation VAR model is widely used in the study of nonlinear dynamic relationship. Because Markov mechanism transformation can well depict the process of nonlinear mechanism transformation of time series, however, there are many problems in the model, such as high degree of dependence on data, excessive estimation of parameters and low degree of fitting, etc. Bayesian statistical inference and the application of MCMC algorithm provide a good solution to these problems. In this paper, the Bayesian Markov transform VAR model is constructed for the modeling of stochastic parameters in the process of analyzing the relationship between stock market returns and inflation volatility. Then, by analyzing the statistical structure of the model and setting the prior probability distribution of the parameters, the Bayesian statistical method is used to infer the posterior distribution of the model parameters, and the corresponding two-time Gibbs sampling algorithm is designed to estimate the model parameters. Finally, we use Bayesian Markov transform VAR model to study the dynamic relationship among stock return rate, inflation fluctuation component and inflation trend component. The results show that the Bayesian Markov transform VAR model can more accurately describe the nonlinear dynamic characteristics of the volatility relationship between three variables. In the specific process of mechanism transfer, stock yield is positively correlated with persistent inflation rate, temporary inflation rate in "market contraction zone" system, and weakly positive correlation with inflation rate in "market expansion zone system" in the short term. In the long run, there is a negative correlation between Fisher effect and inflation rate, which indicates that Fisher effect and agency effect are reflected in different mechanisms of market. It shows that the fluctuating relationship between stock yield and inflation does not always follow Fisher effect, and there may be a variety of relationships.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F224;F832.51;F822.5

【相似文献】