An innovative high-frequency statistical arbitrage in Chinese futures market

Abstract

The primary use of futures is hedging risk. Traders in the spot market can hedge certain risks through the futures market. With the development of the futures market, the arbitrage transactions around futures have attracted increasingly attention. The aim of this paper is to establish an innovative and unique pairs trading framework, and use it to test the effectiveness of China's futures market. The framework for pairs trading is based on cointegration test, Kalman filter and Hurst index filtering. We use the data of 47 commodities with relatively good liquidity in the Chinese commodity futures market. We apply the representative index of China's commodity futures market, "Wenhua Commodity Index" as the benchmark, to evaluate the performance of the strategy and compare it with the benchmark model. This study found that, according to the pairs trading framework, after considering transaction costs, the cumulative return in the sample reached 81%, the cumulative return out of the sample was 21%. It is worth noting that the out-of-sample maximum drawdown achieved excellent results of no more than 1%. In the same period, trading the Wenhua Commodity Index with a "buy and hold" strategy achieved a gain of 31%, but with maximum drawdown reached 15%. The values of our paper are, first it proves that Chinese commodity futures market is not a weak-form efficient market, because technical analysis based on machine learning could obtain excess returns. Second, this research combines the theory and practice of statistical arbitrage, which also provides guiding significance for investment practice.