Realized GAS-GARCH Model And Correlation Measure Based On High-frequency Data

Precious metals trading refers to the process in which investors buy low and sell high to earn the difference when they are optimistic about the precious metals market.It can also be a hedging method adopted by investors in the absence of optimistic economic prospects to achieve the preservation and appreciation of assets.Since the world's precious metal reserves are certain,precious metals can be used as a tool to preserve value,not only for investing to increase personal assets,but also as a token to play its currency attributes.Precious metals have a good hedging function and can be used to maintain value during inflation.At the same time,precious metals such as gold and silver can circulate all over the world,it is difficult to manipulate their prices in the market and is not easy to cause collapse.Also there is no problem of depreciation.Conducting a 24-hour transaction allows investors to have more investment opportunities.Therefore,mastering the market volatility and investment risk of precious metals plays an important role in the study of financial markets.With the development of technology,the availability of high-frequency data is increasing.Many scholars have conducted volatility research of the realized measure based on high-frequency data.Combined with the research status,this paper firstly uses the traditional GARCH model to carry out empirical analysis to prove the applicability of the GARCH model to financial time series.Then the Realized GARCH model combined with the Realized volatility is generalized.The GAS impulse response function is introduced for the fact that the quadratic impact response function is overreacted.The relaxation exponential impulse response function is also considered,to prove the setting of quadratic impact response function is overreacted.At the same time,it is extended to the case of heavy tail distribution,and the residuals are subject to the standard t distribution and ged distribution respectively to adapt to the characteristics of the peak and heavy tail of the financial time series.When comparing the effects of the tail risk metrics,the VaR effect is used to compare the robustness of each model.This paper also establishes the Realized GAS-GARCH Copula model,using the Realized GAS-GARCH model as the edge distribution function,and uses the tCopula function to describe the correlation structure between financial multivariate time series.The empirical results based on high-frequency data of Au888 and Ag888 futures show that the risk measurement effect of Realized GARCH model is better than that of traditional GARCH model.The effect of Realized GARCH model with different impulse response functions is significantly different,and for different tail risk levels,the difference is not the same.The Realized GAS-GARCH Copula model constructed in this paper is robust to the correlation measure.In general,there is a great correlation between the price fluctuations of gold future and silver future.