【作者】 孙峰；

【导师】 刘澄； 刘祥东；

【作者基本信息】 北京科技大学 ， 企业管理， 2020， 博士

【摘要】 本研究分析了行业资产价格波动的系统性影响因素,指出投资者情绪是影响行业资产价格波动的决定性因素,并对投资者情绪在不同行业中的时序传播特点、规律和空间分布进行了计算和分析。(1)建立MVS三维向量因子模型,分析行业资产价格波动的影响因素体系和影响程度。分析了多个行业资产价格波动的联动效应和运行机理,对影响行业资产价格波动的决定因素机制进行了实证研究,用以检验MVS三维向量因子模型。数据分析表明,投资者情绪变量与行业资产价格波动变动具有显著相关性,相对于宏观经济变量和行业价值变量来说,投资者情绪变量对行业资产价格波动产生着更大的影响,它是影响行业资产价格波动的决定性因素。(2)采用因子降维的方法从行业多变量指标中提取投资者情绪变量。对不同行业多变量进行降维提取投资者情绪因子时,技术难点在于须先将不同行业多变量整体降维至相同坐标轴以便提取投资者情绪因子,然后再分行业利用投资者情绪因子进行分析。采用实证数据对投资者情绪在不同行业中时序传播进行测量和分析,结果表明投资者情绪以锯齿状向前传播时,随时间推移其趋势表现出阶段式反转效应、动量相关效应、群集波动效应、协同运动效应。投资者情绪在同一行业不同时期以及不同行业同一时期均存在着不同的传播速度,速度正负值变化揭示出阶段式反转效应,速度数值变化也表示出协同运动效应。(3)通过有限元热—流耦合技术,构建2D投资者情绪耦合紊流模型,对投资者情绪在行业中传播的空间分布进行计算和分析。信息在投资者中传播的现象形成了“信息环流瀑布”,其速度并不服从均匀分布,这造成了投资者情绪传播平均空间分布并不服从均匀分布。更多还原

【Abstract】 This research analyzes the systemic influence factors for industry asset price volatility and points out that investor sentiment is the decisive factor to affect industry asset price volatility.It also computes and analyzes time series contagion characters and space distribution for investor sentiment in different industries.(1)This research builds MVS three dimensions vector factor model to analyze the influence factors system and influence extent for industry asset price volatility.It also analyzes the linkage effect and running mechanism for asset price volatility in many industries,and gives an empirical research to decisive factor mechanism affecting industry asset price volatility to test MVS three dimensions vector factor model.Data analysis shows that investor sentiment variable is significantly correlative to industry asset price volatility.Comparing with macroeconomic variable and industry value variable,investor sentiment variable has greater influence on industry asset price volatility,which is the decisive factor affecting industry asset price volatility.(2)This research uses factor decreasing dimensions method to extract investor sentiment variables from many variable indicators of industries.To extract investor sentiment factor by decreasing dimensions for many variables in different industries,the technical difficulty is that the dimensions for many variables in different industries should be totally decreased to the same coordinate axis to extract investor sentiment factor.Then investor sentiment factor can be used to analyze according to separate industry.This research uses empirical data to measure and analyze time series contagion for investor sentiment in different industries.The result points out when investor sentiment makes the contagion forward with jigsaw shape,its time tendency shows periodical reversion effect,momentum correlation effect,clustering volatility effect and comovement effect.Investor sentiment has different contagious velocities in different stages for an individual industry and in the same stage for different industries.The positive and negative number changes for velocity show periodical reversion effects,and velocity number change also shows comovement effect.(3)By using finite element thermal-flow coupling technology,this research creates 2D investor sentiment coupling turbulence model to compute and analyze the space distribution for investor sentiment contagion in different industries.Information makes the "information cycle cascade’ contagion among investors.Its velocities are not uniform distribution,and these lead to that space distribution of investor sentiment contagion is not uniform distribution.更多还原