| In recent years,the global financial market has been developing and perfecting with the promotion of artificial intelligence and technological innovation,showing a series of characteristics.On the one hand,the linkage of financial markets in various countries is more obvious.For example,the rise and fall of American stocks often affect the performance of domestic stock market.On the other hand,the rise of quantitative investment has changed the structure of market participants and diversified investment strategies and tools.However,the frequent occurrence of extreme events such as financial crisis and market collapse has caused investors to suffer a lot of losses,and also made people pay more and more attention to the tail risk of financial markets.Traditional mean-variance theory assumes that the distribution of return on assets is normal distribution.However,more and more evidence shows that,for example,in the 2008 financial crisis,the distribution of return on assets is peak and tail,which does not conform to the assumption of normal distribution.In this paper,we study the domestic commodity futures market based on high-order moments.On the premise that the logarithmic return of assets follows stochastic differential equation with jump,we give the relevant limit properties of high-order moments under high-frequency data,which is elegant and concise in mathematics.Then,the optimal demand model of short hedging in commodity market is established,and the optimal hedging ratio of futures under normal distribution and skewness normal distribution is investigated to measure the impact of asset returns which do not have a normal distribution on market hedgers.Finally,we select the historical price data of domestic commodity futures market,construct high-order moments from high-frequency data perspective and low-frequency data perspective respectively,and investigate their historical performance and market characteristics.At the same time,we use Rank IC test,quantile combination test,Fama-Macbeth regression and other methods to examine the explanatory ability of high-order moments to expected returns of commodities. |
