Futures Trading Margin Model To Evaluate The Optimal Proportion

As an integral part of modern financial markets, future market is playing an important role in the world economy. Futures margin system is the core of futures market risk management system, so completing the margin system has become the key of improving risk management system. China's current futures margin system is the static margin, while most of the futures exchanges of the word is taking dynamic margin system. With the rapid development of China's market economy, how to strengthen risk management of futures markets and increase market efficiency has become an important issue.The purpose of this paper is to find an optimal model for the margin calculation, in which the level of margin can not only reflect the volatility of futures prices dynamicly but also better cover the daily risk of price fluctuation. The main content of this paper is as follows.Introducing various currently precious margin calculation model in the world, which include the coefficients of price risk, SMA, EWMA, GARCH, SPAN and TIMS systems, and analyzing the characteristics of each model. Then establish the index system of the margin level evaluation. In addition, under consideration of the present conditions of China, doing empirical research with the Shanghai Futures Exchange copper futures data on the three kinds of models SMA, EWMA, GARCH, and selecting the GARCH model after analysis and comparison. Finally, analyzing the reason of the gap between three model test results and expected results, and proposing method to improve the GARCH model by using RBF neural network,then comparing and analyzing RBF neural network model and GARCH model results.The main conclusion is that with the increasing complexity of three models of SMA, EWMA and GARCH,the model results are getting better and better, without considering the time cost,the GARCH model improved by the RBF neural network is more advanced than the GARCH model. The deficiency is not considering the following factors:market liquidity, the futures price limit restrictions and optimal calculation method of RBF neural network.