A Study On Volatility Characteristics And Forecast Of Internet Monetary Fund Yield

Internet culture has been rapidly infiltrating into China since 2013,a series of new varieties have been derived from traditional financial products.Nowadays,there're not only many kinds of Internet Monetary Found financial products,but also a large number of users and a huge fund size.Characteristics with higher yield,quite strong liquidity,flexible access,and convenient usability make it has a good development prospect,the degree of recognition and acceptability have also been improved.However the yield of Internet Monetary Found is getting lesser and lesser recently,the current situation of strong supervision makes investors unable to grasp market dynamics accurately when making investment decisions.So study the news in the market will have a vital practical significance on what will be affected of the yield of Internet monetary fund,how to better understand the asymmetric characteristics of the yield of Internet Monetary Fund,and how to predict the future short-term yield of Internet Monetary Fund.This article selects the two most representatives of Celestica and Huaxia(Respectively docking with Yu'E Bao and Money HD)as sample products in our Internet Monetary Fund,with the help of Tiantian Fund Website to collect the daily 10 thousand portions of accruals from the two Funds during the three years from the beginning of July 2015 to the end of June as sample data in the historical simulation interval,and divide the interval according to the date of the lowest yield as demarcation point.With the help of descriptive statistical method and GARCH Model to do the research of the volatility of the whole and the staged yield rate sequence on Yu'E Bao and Money HD respectively,and the empirical results show that their rates of yield have similar changes over the past three years,which have a certain volatility with a downward trend,in addition with autocorrelation and ARCH Effect.However ARMA(5,2)Model and ARAM(1,1)Model can respectively solve well the correlation of the yield rate sequence between Yu'E Bao and Money HD,and for the residual sequence text of the mean equation can choose EGARCH Model to build model and analyze on the whole and the staged sequence and then come to a conclusion separately.There's an asymmetry in the yield rate sequence of Yu'E Bao and Money HD,the affect from positive and negative news are different in different time.In order to verify the predictability of yield rate sequence,depend on R/S analytical method to obtain Hurst value to verify the nonlinearity of yield rate sequence.If the result of Hurst value of the yield rate sequence of Yu'E Bao is above 0.5,it will be with predictability.Taking Yu'E Bao which is with long-term memory property as a sample fund to forecast and research,and make use of the daily 100 thousand portions of field to be as sample data in the post-prediction interval in the coming week,then depending on ARMA(5,2)-EGARCH(1,1)Model and Extreme learning machine method to forecast the short-term field.The empirical results show ELM Model has a higher prediction precision than pattern assembly on the short-term field prediction of Internet Monetary Found,and also has a better applicability on models.Finally,according to the research conclusions as mentioned earlier,and in view of Internet Monetary Found which be with wide varieties and serious homogeneity,and the current situation of the bank financial products which be launched continually and supervised tightly,we will make related recommendations from three aspects of the internet enterprise,the investor and the government.