投资者关注对中国黄金价格波动率的影响研究  

A study on the impact of investor attention on Chinese gold volatility

作　　者：梁超 魏宇 马锋[1] 李薇 LIANG Chao;WEI Yu;MA Feng;LI Wei(School of Economics and Management,Southwest Jiaotong University,Chengdu 610031,China;School of Finance,Yunnan University of Finance and Economics,Kunming 650221,China;School of Finance,Southwestern University of Finance and Economics,Chengdu 610074,China)

机构地区：[1]西南交通大学经济与管理学院,成都610031 [2]云南财经大学金融学院,昆明650221 [3]西南财经大学金融学院,成都610074

出　　处：《系统工程理论与实践》2022年第2期320-332,共13页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(72071162,71971191,71701170);云南省高校科技创新团队项目(201914);云南省科技计划基础研究重点项目(202001AS070018)。

摘　　要：黄金具有商品和货币的双重属性,是投资者进行资产保值及增值的重要手段.本文从行为金融理论出发,采用广义自回归条件异方差混频数据抽样模型(GARCH-MIDAS),探究百度指数和谷歌趋势对中国黄金价格波动率的预测能力.同时,引入了全球经济政策不确定性(GEPU)指数以及地缘政治风险(GPR)指数变量,检验对黄金波动的影响.进一步地,使用模型信度集(model confidence set,MCS)和方向检验(direction-of-change,DoC)两种评价方法检验各模型的样本外预测精度.实证结果表明,谷歌趋势能够显著大幅提升中国黄金价格波动率的预测精度,为准确预测我国黄金波动提供了新的视角,为稳定金融市场提供了可靠保证.Gold has the dual property of commodity and currency,which is an important means for investors to maintain and increase the value of assets.Based on the behavioral finance theory,this paper adopts the generalized autoregressive conditional heteroscedasticity mixed data sampling model(GARCHMIDAS) to explore the predictive ability of Baidu index and Google trend to the Chinese gold price volatility.At the same time,global economic policy uncertainty(GEPU) index and geopolitical risk(GPR) index variables are introduced to test the impact on gold volatility.Furthermore,two evaluation methods,model confidence set(MCS) and direction-of-change(DoC),are used to test the out-of-sample prediction accuracy of each model.This study employs the closing price data of Au(T+D) gold deferred trading contract with large trading volume as the research object.The sample interval from January 4,2011 to December 31,2019,and a total of 2,186 daily data are obtained from the CSMAR database.The monthly data of Baidu index is from http://index.baidu.com,and the monthly data of Google trends is from https://trends.google.com/trends/.We obtain the monthly data of GEPU and GPR from http://www.policyuncertainty.com.Based on the empirical results of the MCS and DoC tests,the GARCH-MIDAS-Google model performs the best predictive power than other competing models,which means Google trends contain more useful information for the Chinese gold price volatility.Moreover,our results are robust to different forecasting window.Therefore,our findings provide a new perspective for predicting China’s gold volatility,provide a reliable guarantee for stable financial stability,and provide valuable information for policy makers.

关 键 词：波动率预测 黄金价格 GARCH-MIDAS 百度指数 谷歌趋势 

分 类 号：F832[经济管理—金融学]