Research Of Short Selling Trading Strategies Based On Copula Theory

In1970s, Quantitative investment began to rise in the United States capital market. Benefit from its many advantages, quantitative investment quickly became mainstream investment patterns of the USA. Among the many quantitative investment investors, Simmons is most successful, whose medallion fund creates the myth that it won the annual average net return of35%(net of fees) in20years continuously and stably. Moreover, China’s quantitative investment is still in its infancy, but with the gradual improvement of China’s capital market, quantitative investment will have increasingly bright prospects. Meanwhile, the development of quantitative investment can reduce speculation on the market, reduce market bubble, in turn, and promote capital market development and improvement.Under this background, we want to establish a quantitative investment strategy model based on Copula theory, so we can apply it to the capital market of our country. This paper focuses on the establishment of selling short, so that investors will have more investment strategies to selection (compare to traditional terms buy more), to help investors obtain excess returns.This paper selects the high frequency data of soybean oil index and palm oil index in2012September to December a total of four months as the research object. The research object is weighted average of main contract on market, index can reflect the continuity of research objects, so soybean oil index and palm oil index are the suitable object of our study.The investment strategy in this article is mainly based on the tail correlation of the Copula function. With the high lower tail correlation, if soybean oil fell, then the probability of palm oil fell is very high. That the palm oil index does not fell in reality is a "non-normal" circumstance in the tail correlation of the Copula function. We believe that the palm oil index will fall, and this time is an investment opportunity. Known from the definition of Copula function, Copula function is the connection function to describe the marginal distribution of two random variables. To ensure the success of the investment strategy, we need to describe the marginal distribution of the two group objects accurately. The more accurate description of the marginal distribution, the more accurate fitting of Copula function, and the more accurate investment opportunities that may arise. As the distribution of the high-frequency data in the financial assets do not have the normality, and generally have "Rush fat tail" nature, in order to more accurately portray marginal distributions of the objects, this article also abandoned the traditional way to assume the Returns follow a normal distribution, and shift to kernel density estimation, the non-parametric estimation method.To test the actual data by using the research method, our paper obtained a very good accumulation rate of return from September to December, the cumulative rate of return are17.70%、3.45%、8.97%、-0.82%,which belongs to the excess rate of return in today’s financial investment theory. The numbers of transactions were96、57、79、6, the total number reached to238, belonging to typical high-frequency trading, form September to December four months in the single largest loss was-1.4699%. Obviously the loss rate is acceptable to us. In this paper, the empirical results show that the investment strategy have certain reference significance.The innovation of this paper has three main points:first, the paper break investors’inertial thinking, focusing on short selling transactions; second, creatively use Copula tail correlation structure to struck trading strategies; third, because the large amount of data and calculation, In the process of parameter fitting to obtain the optimal combination of the parameters, quickly get optimal parameters using genetic algorithm, to avoid the shortcomings of time-consuming exhaustive algorithm.The two inadequacies of this article are as below:First, the ability of Copula fu nction to describe the lower tail correlation is much higher than its ability to describ e the upper tail correlation. Therefore, investors can get a poor buy signal, which lea ds to the loss of some profits. Besides,the lack of precision in upper tail correlation may reduce the investment rate of return. Second, unacceptable losses did not appea r in the empirical process. But in the excessive reliance on program trading and with out thorough consideration circumstances, once a very special situation occurs, it may result in a great loss.