Empirical And Trading Strategies Study Of CSI300Stock Index Futures Using VPIN

With the rise of quantitive investment, high frequency trading plays a much more important role in foreign markets. In China, thanks to the T+0exchange rule of Stock Index futures market, high frequency trading also gains its popularity.Base on the concept of Probability of Informed trading (PIN), Easley (2012) use volume time to replace physical time and propose what they call Volume-Synchronized Probability of Informed trading (VPIN), which does not require the intermediate estimate of non-observable parameters. They claim that VPIN is a useful indicator of short-term price volatility. Thus we try to design trading strategies base on VPIN.We use high frequency trading records of CSI300Stock Index futures from Nov.2011to Oct.2012. VPINs and returns are calculated within a15minutes time frame. In order to make sure that VPIN has the information for future returns, we first employ a threshold regression model for empirical study. According to the method proposed by Hansen (1996) for estimating parameters and testing the effect of threshold in an threshold model, we regress future returns on its first-order lags and take VPIN as threshold variable. We find significant threshold effects and parameters. There is negative relation between return and its first-order lag when VPIN is large. The result confirms that VPIN do play a role in forecasting future returns.Base on the empirical results, we design three trading strategies. Strategy I only contains the basic key rule:buy or hold a long position when VPIN is high and current return is negative, sell or hold a short position when VPIN is high and current return is positive. In addition to Strategy I, Strategy II incorporates the mechanism of loss limit and load up. Strategy III adds the moving average indicator to Strategy II. With some proper parameters, all the strategies behave well in back testing. Specifically, Strategy I has the best results but also is more risky. Strategy III is medium and strategy II the worse. In conclusion, it might be profitable to design trading strategy based on VPIN.