Pairs Trading Analysis Of ETF And Stock Index Futures Based On 1-Minute Transaction Data

In the rapid development of Chinese stock index futures market and the global ETF market,it is particularly important to analyze on the application of the pairs trading strategy to the Chinese ETF and futures markets.By applying the pairing trading strategy based on the distance method and the cointegration method between China's ETF and the stock index futures market,oneminute high-frequency transaction data of ETFs and stock index futures corresponding to the HS300 index and the Shangzheng50 index are selected respectively to test this strategy.The strategies have significant effectiveness in the Chinese market and can obtain high yields at low risk level.While recognizing that pairs trading's profitability,it is also important to note that transaction costs can bring a negative impact on it.In addition,in terms of open and close thresholds,the lower the open threshold,the higher the yield level of the strategy,and the change in the close threshold has no significant effect on the strategic return,which also means that Chinese ETF and stock index futures market have strong mean-reversion characteristic.In terms of the length of the pairs formation period,it should not be too long or too short because the selection affects the description of the historical average level,which in turn affects the profitability of the strategy.In the pairs trading process of distance method and co-integration method,the pairs of ETF and futures of current month always gain higher yields.The phenomenon that stock index futures with different maturity have obvious influence on the yield of the transaction may originated from that futures with a long maturity may bring more space to the implementation of the statistical arbitrage strategy because of the greater possibility of price fluctuations,but the positive yields they gain cannot be covered by the negative yields caused by weaker co-integration relationship.Ultimately,futures of current month become a better product in traditional pairs trading.By comparing the difference between the yields obtained by the distance method and the co-integration method,co-integration method is better than the distance method.In the period of rise,yields are similar,but during the period of decline,co-integration method superior to the distance method.That is to say,when the stock market shows a downward trend,strategy based on co-integration method has better profitability.In addition to the traditional methods,this paper uses the support vector regression method in machine learning technology to predict the price of stock index futures and ETF to find arbitrage opportunities.This method can obtain better yield with less training samples,reflect the current price change situation in the market quickly,and this method improved the strategy adaptability in stock index futures with long-term maturity.In the comparison of the three methods,the traditional strategy based on distance method and co-integration method are less risky and can obtain relatively stable positive returns.The risk of support vector regression method mainly comes from the period of stock market goes down,but the latter has a higher expected return with risk of fluctuation to some extent.