Empirical Study On Time Interval Between Extreme Volatilities Of Futures Return Rate

Extreme volatilities of financial assets prices have two very important features:the price is discrete in space-dimension,and in the time dimension, the time interval between two contiguous extreme price volatilities is not equidistant.Therefore, we should study the financial extreme volatilities from the two dimensions, However, foreign and domestic scholars make extensive research on the extreme thick tail of financial assets return rate in space-dimension. But very little attention to the laws of the extreme volatility in the time dimension.This paper selects futures contracts of copper in SHFE, hard wheat in ZCE and soybean in DCE as samples, firstly, applys extreme value theory to analysis thick tail characteristics of commodity futures huge fluctuations,capturing the dimensional feature on extreme volatility of futures returns; and then, uses methods studying the time interval of extreme events in weather,earthquakes for reference,combining theoretical derivation and empirical test, analysis the statistical characteristics of time interval between extreme volatility, studys the time interval distribution characteristics of extreme volatility in futures return rate, puts forward prediction and application methods for the time interval between extreme fluctuations. While, takes advantage of autoregressive conditional duration model to quantitative description sequence correlation of the interval between extreme volatility in futures, provide an effective way to predict the interval between concentration occurred extreme fluctuations. And further analysis the existence "time effect" about extreme volatility of copper futures. Through research, obtaine the following conclusions:(1) The time interval sequence of Futures extreme fluctuations has spike and thick tail characteristics, the time interval of futures extreme fluctuations subject to generalized extreme value distribution.(2) It is feasible that applying the conditions expection-fluctuations of generalized extreme value distribution to predict extreme volatility interval.(3)The time interval sequences of futures extreme volatility have self-relevance, ACD model can be used to quantitatively describe the self-correlation(4)There are some inspiration phenomenon from time analysis futures extreme fluctuations...