| In March 2010,China’s securities market officially launched the margin financing and securities lending business,which marked the formal establishment of the short-selling mechanism in China’s securities market.The establishment of the short-selling mechanism is constantly promoting the development of China’s financial market,and also making the statistical arbitrage trading strategy.It has achieved certain development in China’s financial market.As a market-neutral investment strategy,the statistical arbitrage strategy can be free from the influence of market conditions,allowing investors to trade in a market under any market,by buying stocks whose relative prices are undervalued and selling relative prices.Overvalued stocks to get a certain profit.And statistical arbitrage can also reduce the risk to a certain extent,and increase the benefits under the premise of low risk.In addition,statistical arbitrage can also promote a virtuous circle of capital,improve the efficiency of capital markets and ensure the sound operation and coordinated development of the capital market.When establishing the statistical arbitrage model,we must first screen out a set of highly correlated paired assets,and conduct statistical analysis on the historical price data of the set of assets,using the unit root test,cointegration test and error correction model to the group of assets.A cointegration model is established to determine the tradability and transaction ratio of this group of assets.Then,the sequence of the price difference obtained by the cointegration model is analyzed to obtain the trading rules.Trading rules typically include trading triggers and trading stop loss signals.Since the price difference sequence of the paired assets has the property of mean return,that is,the price difference sequence may deviate from the mean value for a short time,but the market adjustment will return to the mean level at some later time,and the investor can perform the time in the mean deviation time.Arbitrage gains.When there is no return trend in the case where the spread sequence continues to deviate from the mean level,the risk should be immediately stopped to control the risk.When domestic and foreign scholars conduct research on the determination of statistical arbitrage trading signals,most of them extend the classical fixed parameter method and establish many novel statistical models.In the empirical study of statistical arbitrage strategy for banking stocks,this paper first uses the distance formula to select the group of stocks with the highest degree of correlation,and then uses the cointegration theory and error correction model to determine the trading ratio of the paired stocks and obtain the spread.The sequence,followed by three methods to calculate its trading signals:Firstly,using the fixed parameter method simulated on the sample historical data to determine the optimal trading signal;Secondly,considering the variance of the financial time series,using the GARCH model The time-varying variance of the sequence of spreads is estimated to determine an optimal trading signal;Finally,the O-U process is established to describe a sequence of spreads having a mean recovery characteristic to determine an optimal trading signal.The transaction signal is used for trading to calculate the specific income.And through arbitrage back analysis of the data outside the sample,the following conclusions are obtained:(1)From the perspective of profitability,the fixed parameter method has great instability,while the GARCH model and the O-U process method have moderate benefits but show high stability.It shows that the trading strategy based on GARCH model and O-U process method is a relatively conservative arbitrage method,while the fixed parameter rule is suitable for some investors with risk preference to make investment attempts.(2)From the point of view of stop loss rate,the stop loss rate of the O-U process is very high,indicating that there may be a higher risk,and the stop loss rate of the fixed parameter method and the GARCH model method is an acceptable level.It shows that when the return rate is equivalent,the trading strategy based on the O-U process method is more suitable for investors who want to obtain relatively stable income but can also accept appropriate investment risk,and GARCH model rule is more suitable for risk averse who pursues stable income.(3)From the point of view of the number of arbitrage and the number of single arbitrage transactions,the number of arbitrage times of the trading strategy based on the GARCH model method is slightly higher than that based on the fixed parameter method and the O-U process model method,and the number of single trading days is relatively shorter.Considering the short-term arbitrage method,the longer the trading time,the higher the risk.Therefore,the GARCH model is also a more stable and less risky method of determining trading signals. |
PDF Full Text Request