A Study On Fractal Characteristics Of Several Agricultural Product Futures Prices In China

In order to increase the understanding of the characteristics of agricultural futures prices,the paper uses the fractal market hypothesis to study several typical agricultural products in my country.The results show that the each variety has obvious fractal characteristics.In order to choose the most representative variety in the entire agricultural futures market,We use turnover,variety,time to market,etc as selection criteria,and soybean meal,corn,sugar,and eggs is finally selected as study objects In empirical research,we obtain the logarithmic yield series by processing the daily closing price of the weighted index of each variety Then we analyze the yield series to obtain the single fractal and multifractal characteristics of each variety.For single-fractal structure:(1)K-S test and histogram statistics were used to judge whether each species follows a normal distribution The research results showed that the yield series of each variety significantly rejected the normal distribution and showed fat tail structure Among them,the most significant difference from the normal distribution was corn.(2)the square residual Q test is used to determine whether the yield series of each variety is non-linear,and the chi-square distribution is constructed for the square residual of the yield series The statistical result rejects the linear assumption,that is,the series are non-linear.(3)Use the GARCH model to determine the volatility clustering of each variety The results show that the yield series of soybean meal,sugar,eggs,and corn all have ARCH effects and are respectively subject to GARCH(1,1),EGARCH(1,1),and TGARCH(1,1)under the residual T distribution,and GARCH(1,1)under the residual GED distribution.After fitting each series with the GARCH model,the results show that the aggregation effect of return fluctuation is obvious.(4)For the self-similarity of various varieties,This article mainly compares the closing prices under the daily,weekly,and monthly closing prices of various varieties to judge the self-similarity This result can be clearly observed.(5)The study of long memory is the core of single fractal,which is mainly tested by R/S analysis The article calculates the Hurst index and non-periodic cycle of each variety respectively The results showed that the Hurst index of each variety was greater than 05,that is,they all showed persistence Among them,the persistence of corn is the most obvious,its Hurst value is 0650186,and the persistence of soybean meal is the least obvious Data rearrangement also helps to prove that the Hurst index is valid For the non-periodic cycle,we innovate on the basis of V/S analysis,and use generalized Hurst value to assist judgment The results showed that the periods of corn and sugar were 143 days and 152 days,respectively There are two cycles of eggs,102 and 226 days,respectively Because the generalized Hurst value is not obvious in determining the soybean cycle,we judge the cycle of soybean based on R/S and E(R/S),V/S and E(V/S),and the results show that the cycle is 213 days Finally,we test the validity of the cycle of Each breed by E(R/S),and the results significantly rejected the null hypothesis,that is,the cycle is valid.For Multifractal structure:This article mainly judges by using MF-DFA and multifractal spectrum The calculation results show that the Hurst value of each variety changes with the change of q value,that is,they have multifractal Combined with the multifractal spectrum,it can be seen that the multifractal of white sugar is the most complicated,and corn is the least The analysis of the causes of multiple fractals is mainly realized by rearranging the data The results show that long memory,fat tail distribution,and extreme values are all the causes of multifractals The multifractals of soybean meal and eggs are mainly caused by extreme values The multifractal of corn and sugar is mainly derived from fat tail distribution.In general,this article gives a comprehensive and detailed introduction to the fractal characteristics of major agricultural futures,and also carried out a certain degree of innovation in research methods,such as the square residual Q test(the non-linearity),GARCH modeling(the volatility clustering),the significance test of non-periodic cycles,and the causes of multifractals,etc.In addition,the practical significance of each fractal feature is also introduced in detail.