Research On Gold Volatility Forecasting In Uncertain Environment

As a precious metal with commodity and monetary value attributes,gold reserves show the economic strength of a country and play an irreplaceable role in reducing the balance of payments deficit and stabilizing the exchange rate.More importantly,during the economic crisis and the increase of economic uncertainty,gold is used as a hedging tool to provide investors with certain financial loss protection.In recent years,more and more people are engaged in and researching financial markets,and the trend of gold price and its volatility trend have been widely concerned by investors and researchers.In view of this,it is of certain practical significance to accurately predict the volatility trend of the gold market.The GARCH-MIDAS model proposed by Engle(2013)decomposes volatility into long-term and short-term parts,which can process data of different frequencies respectively,reducing the loss of data information.Therefore,it is widely used in volatility research.Compared with the traditional direct embedding of external shock into the long-term volatility equation of the GARCH-MIDAS model,inspired by the study of Jesus et al.(2006),this paper considers the different influences of external shock size on volatility.Based on this,the threshold GARCH-MIDAS-SZ model was constructed after determining the measurement of impact size.Compared with the traditional GARCH-MIDAS model,this type of model retains the advantages of the GARCH-MIDAS model,and further describes the impact of external shocks size on gold volatility.On the basis of considering the impact of external shock size on volatility,this paper further considers the impact of asymmetric behavior of external shock size on volatility,and the asymmetric threshold GARCH-MIDAS-AL-SA class model is constructed,This model examines the impact of the asymmetry of shock size on gold fluctuation.Moreover,the paper presents the construction process and parameter estimation process of GARCH-MIDAS-X model and extended model(GARCHMIDAS-SZ model and GARCH-MIDAS-Al-SA model)incorporating external shock.In addition,a large number of studies have shown that macroeconomic uncertainties such as economy and politics have a significant impact on financial market volatility.In this paper,five commonly used uncertainty indices are introduced into the mixed GARCH model considering shock difference and applied to the volatility research of London gold market.Out-of-sample R-square and Do C tests were used to evaluate the out-of-sample forecasting ability of the model.The in-sample and out-ofsample forecasting results show that both the GARCH-MIDAS-SZ class models and the GARCH-MIDAS-AL-SA class models significantly improve the in-sample fitting ability and out-of-sample forecasting ability of gold market volatility.Among them,the GARCH-MIDAS-AL-SA model considering the large shock asymmetry performs the best.Finally,the prediction results of the model are further confirmed by different robustness analyses.In this paper,GARCH-MIDAS-SZ class models and GARCH-MIDAS-AL-SA class models are used to research the impact of uncertainty index shock size on gold volatility,which provides a new idea for subsequent scholars to research volatility,and enriching the literature on the impact of uncertainty index on gold volatility market.At the same time,the extended model proposed in this paper can be further extended to other financial markets(such as stocks,oil)for research.