Study Of Trading Strategies In Options And Stock Index Futures Based On Stationary Process

After a long period of development,options and futures have become the most widely used derivatives in the financial market.There are numerous trading strategies for options and futures,among which the statistical arbitrage strategy based on the stationary process is an important part.A good trading strategy not only needs high returns,but also needs stable returns.When encountering black swans in the market,losses are less,especially in the options and futures markets.Due to the large leverage,more attention should be paid to risk control and the stability of the strategy.At present,the research on the statistical arbitrage strategy of options and futures focuses on how to improve the return rate of the strategy,and there is little research on the risk control of the strategy.The risk control and stability problems of statistical arbitrage strategy based on options and futures are studied under the framework of stationary process.On the basis of the basic strategy of selling the put option based on the stationary process,the model of hedging strategy and continuous downshift strategy are proposed.They can effectively control the strategy net value reversion during the financial crisis.Then we study the application of mean-reversion process in high-frequency futures data trading,which improves the stability of the strategyIn Us stock market,the long-term annualized income of three index ETF(QQQ,SPY,DIA)are ?>r>0.By taking advantage of the stability of logarithmic rate of return,adjusting the share and delivery price of the put option,a statistical arbitrage strategy based on the stable process of selling the put option can be constructed to obtain good returns in the long run.However,when the market encounters black swan time,there will be a large net reversion of this strategy.So we studied a selling put option hedging strategy based on the stationary process,which is:sell a put option,in the meantime,buy another put options with a lower deliverable price.According to the different deliverable time can be divided into symmetric hedge strategy and asymmetric hedge strategy.The symmetric hedging strategy is:sell a week put option with the delivery price K1 and buy a week put option with the delivery price K2,where K1>K2.According to the choice of K2,we construct four strategies:sell week flat value option,at the same time,buy a put option with deliverable price K2=0.99,0.98,0.97,0.95.For these four strategies,we conducted empirical analysis on the three America stock index ETF's Black-Scholes option of all-time data,the Internet bubble data and the sub-prime crisis,and the real data.We found that most of the K2 hedging options used for hedging were wasted since there were very few sharp falls in U.S.stocks.Thus the long-term return of symmetric hedging strategy is worse than the base strategy,but significantly superior to the base strategy during the two financial crises.American market falls are few and unpredictable,so the hedging strategies must balance both the cost and the effect.We observe that the price of put option with very low delivery price and long delivery price is cheap,but can rise sharply when the market drops a lot.So we propose the asymmetric hedging strategy sell a week put option with the delivery price K1 and buy another put option with very low delivery price K2 and long delivery price.In order to maintain the effectiveness of the hedge options,we stipulate that when the price of ETF rise or decline more than the? or the rest of the time come to a fixed number,sold the original hedge options,then buy new hedge options.We select K2=0.85,0.8,0.75,?=15%,and hold the position for 90 days.Thus for these three strategies,we conducted empirical analysis on the three America stock index ETF's Black-Scholes option of all-time data,the Internet bubble data and the sub-prime crisis,and the real data.We found that,in the long run,the return of asymmetric hedging strategy is similar to that of the base strategy,but it has a significant effect on controlling the retracement.Although it cannot avoid losses,it can reduce the retracement greatly compared with the base strategy,especially during the two financial crises.And we test the stability for all strategy returns,and the results are stationary.The move down strategy is another effective way to control the pullback.The price of put option is a monotonically decreasing function of the underlying price.When the market fall down,we can use a single down position strategy:when the underlying price goes down,the original option price goes up,while the underlying price falls by a certain amount,buy back the original option,bear part of the loss,and sell a new put option which delivery price lower than the current underlying price,so as to reduce the total loss However,we know that the U.S.stock market has the characteristics of slow rise and fast fall.When the market drops sharply,it tends to fall quickly and sharply.Thus when the market decline is large,the strategy that only move once can not be enough control of retracement.In this chapter,we study the strategy of moving positions continuously every time the underlying price falls by a certain degree,move down the position once We conducted empirical analysis on the three America stock index ETF's Black-Scholes option of all-time data,the Internet bubble data and the sub-prime crisis,and the real data.We found that due to the situation of a big drop in a single week is limited,there is little difference between the strategy of move once and the strategy of multiple moves,but in the case of a big drop in a week,the strategy of multiple moves is better than the strategy of move once,also much better than the basic strategy and holding the underlying ETF.And we test the stability for all strategy returns,and the results are stationary.The statistical arbitrage strategy can be constructed after the technical index is stabilized.On this basis,we propose a more stationary stochastic process,called Mean-reversion process,which can be written as the difference of a stationary process.We found that Mean-reversion process converges faster and more stable.Two methods of judging mean process are given:1.by the definition 2.testing the stability of the cumulative sum of the random process.Specifically,under the 5s and 20s data of CSI 300 index futures,we conducted statistical tests on some commonly used stationary technical indicators,such as the stationary MACD,BIAS and the bollinger band.It was found that both the smoothed MACD and BIAS were mean-reversion processes under these two kinds of data,while the smoothed bollinger band was not mean-reversion processes.Finally,in order to explain the advantages of mean-reversion process,we take BIAS as an example,and found that BIAS was not a mean-reversion process under 15min data.On this basis,we improved BIAS to make it a mean-reversion process.It was found that all indexes of the improved BIAS were better than those of the BIAS under the same conditions,which indicating that the mean-reversion process was more stable.