【作者】 王莉莉；

【导师】 吴鑫育；

【作者基本信息】 安徽财经大学 ， 金融学， 2020， 硕士

【副题名】基于已实现EGARCH-MIDAS模型的实证

【摘要】 股票市场作为预测宏观经济变化的“晴雨表”,是一国经济的重要组成部分,在融通资金、资源配置和规避风险等方面起到了至关重要的作用。但是我国的股票市场由于内在机制和规章制度的不完善,在一定程度上需要国家进行相应的宏观调控来防止股票市场的大幅度波动,因此本文研究了宏观经济对股票市场波动性的影响,这对于股票市场的健康发展是非常重要的。传统的波动率计量模型在研究宏观经济变量对股票市场波动性的影响时大多采用相同频率的数据进行建模,这极大地损失了高频数据所包含的有效信息。Engle提出的GARCH-MIDAS就解决了不同频率数据建模的问题,而本文在GARCH-MIDAS模型的基础之上考虑了高频金融数据和波动率的杠杆效应,提出了已实现EGARCH-MIDAS模型,该模型将波动率分解为长期和短期两个成分,将宏观经济变量作为长期波动率的影响因素,分别从宏观经济变量的水平值和波动率两个角度建立单因素和双因素的已实现EGARCH-MIDAS模型研究宏观经济变量对股票市场波动性的影响。在实证研究中,首先讨论了引入高频数据和波动率杠杆效应的已实现EGARCH-MIDAS模型对于股票市场波动性的预测效果,其次分别从宏观经济变量的水平值和波动率两个方面建立单因素已实现EGARCH-MIDAS模型研究宏观经济变量对股票市场长期波动率的影响情况,最后研究包含高频金融数据和低频宏观经济数据的双因素已实现EGARCH-MIDAS模型对股票市场长期波动率的影响。本文实证分析所选取的股票数据是上证综合指数的收益率、基于每5分钟高频数据构建的已实现波动率,宏观经济变量选取了货币供应量、消费者价格指数、银行间30日同业拆借利率、工业增加值和宏观经济政策不确定性指数。根据实证结果表明:单因素已实现EGARCH-MIDAS模型中货币供应量和消费者价格指数在水平值和波动率两个方面对股票市场长期波动率的影响都是非常显著的,而宏观经济政策不确定性指数仅在水平方面对股票市场的长期波动影响显著,银行间30日同业拆借利率仅在波动率方面对股票市场的波动有显著影响,工业增加值在水平值和波动率两个方面对股票市场波动的影响都很小;宏观经济变量的波动率对股票市场长期波动的影响程度要显著高于其对应的水平值;包含高频金融数据和低频宏观经济数据的双因素已实现EGARCH-MIDAS模型在样本内的拟合效果和样本外的预测能力两个方面均优于单因素已实现EGARCH-MIDAS模型。更多还原

【Abstract】 As a barometer to predict macroeconomic changes,stock market is an important part of a country’s economy,which plays an important role in financing funds,resource allocation and risk aversion.However,due to the imperfection of the internal mechanism and regulatory framework of the stock market in China,to a certain extent,the country needs to carry out corresponding macroeconomic regulation to prevent the large-scale fluctuation of the stock market.Therefore,this paper studies the impact of macro-economy on the volatility of the stock market,which is very important for the healthy development of the stock market.Traditional volatility measurement models mostly use the same frequency data to model the impact of macroeconomic variables on the volatility of the stock market,which greatly loses the effective information contained in high-frequency data.In order to solve this problem,Engle proposed the GARCH-MIDAS model based on the MIDAS regression model.This article considers the leverage effect and frequency data,put forward the realized EGARCH-MIDAS model,and decomposed the volatility into two parts: long-term and short-term,and take macroeconomic variables as the influencing factors for long-term volatility.The single-factor and double-factor the realized EGARCH-MIDAS models are established respectively from the level value and volatility of macroeconomic variables to study the impact of macroeconomic variables on stock market volatility.In the empirical research,we first discuss the effect of the realized EGARCH-MIDAS model with high-frequency data and leverage effect on the prediction of stock market volatility.Secondly,the single-factor realized EGARCH-MIDAS model is established from the level value and volatility of macroeconomic variables to study the influence of macroeconomic variables on long-term volatility of stock market.Finally,we study the influence of the two-factor realized EGARCH-MIDAS model including high-frequency financial data and low-frequency macroeconomic data on the long-term stock market volatility.The stock data selected for the empirical analysis in this paper are the return rate of the Shanghai composite index,the realized volatility constructed based on high-frequency data every 5 minutes,and the macroeconomic variables selected are the money supply,consumer price index,the 30-day interbank lending rate,industrial value added and the uncertainty index of macro-economic policies.The empirical results show that the level value and volatility of the money supply and consumer price index in the single-factor realized EGARCH-MIDAS model have a very significant impact on the stock market volatility,and the uncertainty index of macroeconomic policy has a significant influence on the long-term fluctuation of the stock market only in the aspect of level,the 30-day interbank lending rate has a significant influence on the fluctuation of the stock market only in the aspect of volatility,and the industrial added value has little influence on the fluctuation of the stock market in the aspect of level and volatility;the impact of the volatility of macroeconomic variables on the long-term volatility of the stock market is significantly higher than its corresponding level value;and the two-factor realized EGARCH-MIDAS model,which includes high-frequency financial data and low-frequency macroeconomic data,is superior to the single-factor realized EGARCH-MIDAS model in both the in-sample fitting effect and the out-of-sample prediction ability.更多还原