| Openness has become one of the most frequently mentioned words in Chinese capital market in recent years.Projects such as Shanghai-Hong Kong Stock Connect,Shenzhen-Hong Kong Stock Connect,the Bond Pass and Shanghai-London Stock Connect have successively landed.The Chinese stock index has also been included in the MSCI emerging market Index baskets and the FTSE Russell Index baskets,which has brought a large amount of incremental money to the A-share market in mainland China,and has also brought a large number of investors from developed capital markets.Foreign investment has great impact on the investor structures and trading habits in A-share market.For example,the A-share market now pay more attention to the company’s financial quality rather than speculating on single stock.The performance of blue-chip stocks gradually surpassed the performance of the small cap companies in recent years.Among all those imported words in capital market,Programmed trading and Highfrequency trading are undoubtedly the most controversial but also the most eyecatching words.Most of the historical research believes that the use of high-frequency trading and high-frequency data outweighs the disadvantages.In short,a lot of research indicates that high-frequency trading reduce transaction costs,accelerate the process of information reflection in price discovery,and provide sufficient liquidity for the market.Order book data is frequently used in high-frequency trading strategy because it reflects information that was not revealed by low-frequency data.According to Chordia(2000),the order imbalance,as the supplement of volume,is one kind of indicator from order book that can measure the imbalance of the trading power of a stock in a certain period of time.Chordia believes that the order imbalance can reflect the trading behavior of investors.In our research,we use the high-frequency trading data of SSE 50 stocks from June 3,2014 to December 31,2015 to test the impact of order imbalance on stock returns.Due to the characteristics of high-frequency data itself,we performed two important pre-processing on the data before conducting the research:First,we distinguish three time intervals in accordance with the performance of the Shanghai Composite Index.Second,we divide the high-frequency trading data into large,medium and small order imbalance according to the relative trading volume.The division can help us distinguish the types of different traders in the market,that is,the trading behavior of institutional investors,medium-force investors and Small investors are represented by different types of transaction orders.We carried out a regression analysis on the autocorrelation of the imbalance of order flow.The results show that the first-order autoregressive coefficient of order imbalance is significant.We use factor analysis framework suggested by Fama and French to analyze the relationship between order imbalance and future returns.The result shows that order imbalance can significantly influence the current and future returns.We use the three-factor model proposed by Fama and French to test the robustness of the order flow imbalance factor.When order imbalance factor is included in the original model,the adjusted R-squares is improved and the parameters are all significant.We also use threshold regression model proposed by Hansen to analyze the existence of thresholds of order imbalance.The regression results show that if the lag-order order imbalance is used as the explanatory variable,on average,a part of stocks have at least one threshold.Finally,we construct a one-factor trading strategy based on the order imbalance.The backtest results in the sample interval indicate that the accumulated return of some strategies is significantly higher than the benchmark accumulated return. |
PDF Full Text Request