Index and futures affairs monitoring 15-S rate of return calculation, using logarithmic difference. Not including the reward of staying up late. The data is divided into two sub-samples. The first sample, sixteen samples, from 65438+1October 3 to1June 23, 997, including 1 19 trading days, 140 1 year, each. The second example, from June 24th 1997, 65438+February 23rd, but excluding 65438+1October 27th and 28th, includes126th trading days, because of extreme fluctuations and problems in the data records of gene Taq. For more information about these two days, including arbitrage, see Ross and Sofianos (1998).
Table 1 shows that the average value, standard deviation and forindex provided since the maximum lag of 4 hectares and the monitoring income of futures transactions are mpe.
As expected, the fluctuation of the midpoint index (0.00637 and 0.00763) is lower than that of the fluctuation-trading index (0.008 18 and 0.00843) due to the elimination of the bidask rebound. The smaller difference in group (2) is reflected in the reduction of the spread, so it is required to be returned when bidding. In addition, the free index rate of return is significantly greater than the midpoint index. As shown in the fourth list 1, before the implementation, the first-order autocorrelation of the sixtieth second, the trade price index is 0.3 14, and the first-order autocorrelation of the midpoint index is 0.678. According to the corresponding statistics, after implementation, one sixteenth of them are 0.560 and 0.756 respectively. This difference is predictable. Observing the composition of the index from the industry, it is not only a negative component that brings back the bid request and a positive component that causes the return target index, but the midpoint index does not contain negative components. This explanation shows that the existing research may underestimate the positive autocorrelation in index returns.
The volatility of futures is much higher than the index. Because of the return of buying and selling, the fluctuation of futures will overestimate the fluctuation of arbitrage trading. The fluctuation of the index will underestimate the fluctuation, which is due to the positive return of the current index. Miller and others. (1994) provides some theoretical results to support this proposition under a simple AR (1) framework. Miller, muthuswamy and Huali also reported that mpe refers to reduction, regardless of the existence of arbitrage opportunities. Statistics in the table 1 contain all opinions, most of which are not current arbitrage opportunities.