| dc.description.abstract | The purpose of this experimental study was to examine arbitrage mechanism of Taiwan50 index and Taiwan 50 index futureses , regard Taiwan 50 indexes as the target, from study the issue day for the first time of Taiwan 50 indexes of since July 1 , 2003ing, until the track effect of ’’ the optimal investment portfolio of Taiwan 50 index simulations ’’ in this time slot on December 31 , 2004.
In the article applied ’’ Mean-Variance Analysis’’ , through distinguishing the rate of returns and risk of stock and the whole investment portfolio, to build constructing ’’ optimization portfolio of index tracking simulation . Because a risk of making the investment portfolio , the becoming different altogether and counting of the ones that depend on the investment portfolio’s covariance, but not the average of the specific risk, the stock of a group of high risk may still build the investment portfolio which construct out a low risk . And study index duplicating , parameter design , model and using , and in order to compare, in order to moveing window simulation method examine core prediction ability of four association (building during constructing different and when holding ) at the same time. Direct against the question that the optimization portfolio of the simulation index is followed the trail of , use the planning two times to perform the algorithm (quadractic optimization programming ) to ask and solve the optimization under the enviroment of MATLAB7.01.
1. The ’’quadractic optimization model’’ and ’’ industry’’s restriction (stratification) + the quadractic optimization model’’ is no matter in the sample or the track result out off the sample is pretty good .
2. The ’’quadractic optimization model’’ and ’’ industry’’s restriction (stratification) + the quadractic optimization model’’ statistics assay result and rate of returns difference superior to ’’ quadractic optimization model’’, show that the ability to duplicate ’’ the optimal investment portfolio of Taiwan 50 index simulations ’’ in investment portfolio not screened through the industry is better.
3. Under the situation that other conditions do not change, the longer the time window in-sample, track effect of in-sample is better than track effect of out-of-sample. And track effect in-sample that the better the result is, the better the behavior outside-of-sample is.
4. In this studying, the rate of returns of ’’ the optimal investment portfolio of Taiwan 50 index simulations ’’ is slightly worse than the rate of returns of Taiwan 50 futures indexes, but the liquidity of ’’ the optimal investment portfolio of Taiwan 50 index simulations’’ is quite outstanding , have supported this research to have certain actual using value from the side.
5. In order to accelerate experiment time and improve experiment accuracy, except that the planning model optimization two times is asked and solved , still increase experiment automation and elasticity mould group’’s function in the course of developing systematically, make the whole experiment course totally automized. So cooperates with arbitrage trade mechanism and real-time market trade data, the procedure that this research institute develops can apply to the computer program trade. | en_US |