Research On The Design Of Breeding Profit Index And Statistical Arbitrage

Statistical arbitrage is a low-risk investment method based on the assumption of mean recovery and market neutrality.It mainly constructs the spread sequence by looking for the asset portfolio,uses various arbitrage models,sets the open and close line and stop loss line of the arbitrage spread,opens the position when it deviates from the equilibrium level,closes the position when the spread sequence returns to the equilibrium level,and finally obtains the spread income.At present,the spread sequence is mostly constructed based on the cointegration analysis method,in which the asset portfolio can be determined by the upstream and downstream related products of the entity.This paper takes laying hens as an example to construct profit index of industrial chain and carry out statistical arbitrage analysis with it as spread sequence.This method no longer constructs the spread solely through the probability and statistics principle,but fully considers the active market behavior of the entity enterprises to adjust the business scale according to the profit size to complete the construction of the spread.The proposed method has the characteristics of easy popularization in the industrial chain.In order to verify the effectiveness of the proposed method,this paper took related futures varieties of laying hens breeding enterprises from 2017 to 2019 as research objects,and constructed statistical arbitrage models using cointegration analysis and laying hens breeding profit index respectively.In view of the cointegration analysis,the egg and feed combination（corn and soybean meal 3:1）were constructed,and the correlation test,stationarity test and cointegration test were carried out to establish the error correction model and build the spread sequence.The corresponding statistical arbitrage trading signals are determined by O-U process trading signal model,nonparametric symmetric trading signal model and non-parametric asymmetric trading signal model.Aiming at the arbitrage model of breeding profit index,the factors of breeding profit of laying hens were analyzed,including egg price,corn cost,soybean meal cost,other feed cost and labor cost,and the breeding profit index was constructed.Then,the stationarity test is carried out,and the test results show that the breeding profit has good mean recovery,and statistical arbitrage model can be constructed.Finally,the non-parametric asymmetric trading signal model is used to determine the opening and closing signal.In this paper,out-of-sample backtesting is performed with 2020 data to verify the robustness of the statistical arbitrage model.The in-sample numerical calculation results show that the statistical arbitrage model based on profit index can obtain a positive rate of return and a positive Sharpe ratio,with an annualized rate of return of 37.95% and an annualized Sharpe ratio of12.6137,the most outstanding performance.In addition,the performance of statistical arbitrage model based on cointegration analysis is also acceptable,in which the annualized rate of return of arbitrage based on O-U process is 7.08% and the annualized Sharpe ratio is 5.5824,and the annualized rate of return of symmetric trading signal model based on non-parametric method is 6.22% and the annualized Sharpe ratio is2.1899.The annualized return based on the non-parametric method asymmetric trading signal model is 18.32%,and the annualized Sharpe ratio is 13.3755.From the calculation results of out-of-sample values,the statistical arbitrage model based on profit index and the statistical arbitrage model based on cointegration analysis are both relatively stable.