Quantitative Investment Analysis of Cross Period Arbitrage in China's Soybean Meal Futures Market Based on a Three-stage Threshold Autoregressive Model

CHEN Xin-hua

Abstract：Based on the nonlinear characteristics and mean reversion mechanism of price differences between contracts of the same variety with different delivery months in the futures market,a quantitative trading strategy for cross period arbitrage in the futures market was studied using inventory theory,no arbitrage pricing theory,and three-stage threshold autoregression model.The strategy was backtesting analyzed using daily price data of soybean meal futures contracts on the Dalian Commodity Exchange from 2017 to 2021.Research has found that,firstly,in the long run,the upper and lower threshold values of the autoregressive volatility of the cross period arbitrage portfolio spread of soybean meal futures contracts with a difference of four months between the near and far months are not significant,and the profit situation of cross period arbitrage trading backtesting using it as a judgment of the no arbitrage interval range is also unstable;Secondly,there are significant differences in the values and significance of the threshold for cross period arbitrage portfolio price difference sequences of different time lengths.The dynamic quantitative investment backtesting effect using the threshold value of the 400 day period price difference sequence is better than that of long-term and short-term price difference sequences;Thirdly,overall,the risk control effect of cross period arbitrage is better than speculative trading,but different selection methods of non arbitrage intervals can lead to significant differences in backtesting risk of quantitative investment strategies.