Construction And Empirical Study On Medium And High-frequency Calendar Spread Arbitrage Model In Chinese Commodity Future Market

With the gradual development of China's futures market,the calender arbitrage of futures has become one of the most popular pair trading methods for the market participants.The calendar spread arbitrage is constructed by two futures contracts with different maturities and can provide continuous and stable returns for arbitrager,while can greatly reduce the market risk faced by the portfolio.On the other hand,in the current China's financial market,as one of the few financial products that can carry out intra day trading and also provide shorting mechanism,future market provides a good platform for the application and research of medium and high-frequency trading in China's financial market.Based on China's commodity futures market,this thesis studies the calendar spread arbitrage strategy of middle and high frequency future data(1,5,15 and 30 minutes).Different from the traditional interval of arbitrage-free prices strategy or O-U model strategy,this thesis adopts Kalman filtering algorithm to optimize time-varying coefficients,and obtains standardized residuals.Then set open and close threshold signal and design stop loss strategy.The arbitrage strategy is applied to 20 commodity future varieties,and the complete trading price data of year 2016-2017 are used as samples to carry out the empirical research.And then the backtesting results are detailed discussed from three aspects: different time frequencies,different future varieties and different market trends,to analyze the factors affecting the performance of the strategy.The empirical results show that the calendar arbitrage strategy based on Kalman filtering can get a steady net profit on the 20 commodity future pairs,and the iron ore future pair is the best,the hot rolled coil and rebar future pair have the poorest performance.With the increase of time frequency,the performance of most pairs has been improved.This strategy is stable in different market environments.The smaller the ratio of the transaction cost to the minimum price change,the smaller the number of slipping points,the better the performance of the futures pairs.