Empirical Analysis Of Statistical Arbitrage Strategy

Stock markets are very mature and statistical arbitrage investment strategy is widely used in western countries. It has become a common strategy used in hedge funds and investment banks. China's stock market launched securities margin trading system officially in March 31, 2010. The launch of margin trading not only brought short-mechanism to Chinese securities market and made the Chinese stock market bid farewell to the era of unilateral market, but also provided more development space to the statistical arbitrage strategy.The core idea of statistical arbitrage strategy is establishing a paired stocks. Out of the premise of economic implications, only using rational quantitative methods to structure investment portfolio can further immune systemic risk and get the excess return which does not exist or exit lower risk. The main way of arbitrage is to find the long-term balance relationship of each portfolio. When the spreads of two varieties in a combination deviate to a certain extent, it appears arbitrage opportunities and we begin to open positions. When the spreads return to equilibrium level, we close the position.There are many specific methods in statistical arbitrage strategy. The cointegration strategy and principal component analysis strategy are the main methods. We select the SSE 50 Index and it's constituent stocks' historical closing price data to do empirical analysis in this paper. On the one hand, we use cointegration strategy to do the research. Firstly, selecting relatively high correlation portfolio to do cointegration test one by one, which use E-G test and Johansen test. The portfolio of Bank of Huaxia and Bank of Beijing satisfies cointegration test, so we build ECM(the error correction model) and obtain it's residual series. Then using mean reversion model, AR model, GARCH model to analyze residuals series and comparing three arbitrages' earnings effects. By comparing, the effect of mean reversion model is the best model; On the other hand, taking principal component arbitrage strategy to do the research. Firstly, using principal component analysis to select five main components to represent SSE 50 Index. Then build mean reversion model towards Bank of Huaxia and main component to make arbitrage trading. By comparing the arbitrage income of cointegration strategy and the principal component arbitrage strategy, we find the effect of principal component arbitrage strategy is superior to the cointegration strategy.