Extreme Value Theory Analysis Of Commodity Futures Index Industry Risk

With the continuous development of domestic production and economy,commodity futures have not only increased transaction volume year by year,but also have increasingly rich varieties.Commodity futures,as an important part of the allocation of major assets,has a wide range of commodity futures contracts and different time limits.There are many factors affecting prices.The commodity futures index can not only be used to describe the price trend of commodity futures,but also a good tool for analyzing commodity futures risks.This article uses the commodity futures index as the research object to analyze the risk of commodity futures in different industriesThis paper attempts to analyze and compare the risk characteristics of the commodity futures industry by calculating the risk values of commodity futures indices in different industries.The selected research objects are the composite index of China Securities Industrial Commodity Futures Index and the index of 8 China Securities Industry Commodities(agricultural,metal,chemical,energy,corn,industrial metals,textiles and oils),the sample selection time range is from January 4,2007 to March 2,2018.First of all,the daily closing prices of futures indices for eight industries are used for AR(1)-GJR-GARCH(1,1)model,an independent and identically distributed standardized residual sequence is obtained for each industry,and then the generalized Pareto model is used to estimate the distribution of normalized residual term sequences of various industry futures indices.Carlo simulation of risk values in different periods(one week,one month,one year)and different confidence levels(90%,95%,99%).Based on the risk values,we conducted mutual comparison and analysis among industries and found the futures index of textile industry has the largest simulated risk value,while the oil industry and the industrial metal industry are the industries with relatively low price risks for the two futures indices;then the t-Copula model is used to simulate the eight industry combinations at different time range(one week,one month,one year)and risk values with different confidence levels(90%,95%,and 99%),and compare them with the risk values of the commodity futures composite index,then find that in a shorter period(one week,one month),the t-Copula model can more accurately estimate the combination correlation,but within a longer period(one year),the t-Copula model has a poor estimation of the inter-industry correlation and there is a difference between risk value estimated by t-Copula model and risk value estimated by commodity futures composite index.