The Research of Long-term and Short-term Relationship between Prices and Return Rates of Taiwan Stock Exchange Capitalization Weighted Stock Index Futures and the Spot Markets

• Main Authors: San-Hua Wang, 汪三華
• Other Authors: Su —Ling Lai
• Format: Others
• Language: zh-TW
• Published: 2001
• Online Access: http://ndltd.ncl.edu.tw/handle/60353174258440966455

碩士 === 中國文化大學 === 國際企業管理研究所 === 89 === This research studied the relationship between the Taiwan stock index futures and the spot markets, and compared the predicting ability of the different predicting models. By applying to “the autoregressive model”, “ploynomial distributed lag regression”, and “the generalized autoregressive conditional heteroscedasticitymodel”, the results present: 1. The index series of the stock and the price series of the futures have the non-stationary process, and the undulation of them will change with the time. But the return rate series of the spots and the futures have the stationary process. 2. The index series of the stock and the price series of the futures have the cointegration. The meaning is the two series have the same changing direction,and the range of variation of both will keep within certain degree. 3. By comparing the verifying results, there is no such better predicting model to predict the price of the futures with the stock index. And the AR model is proved to be the better model to predict the return rate of the futures with the return rate of the stock index. 4. The error analysis between the stock index and the price of the futures can fit in all the normality, the independentness, and the homogeneity. The error analysis between the return rate of the futures and the return rate of the spots can only fit in the independentness, and the homogeneity, but it can’t fit in the normality. It shows there are some other relationships in the error. 5. Reviewing the error autocorrelation test and ARCH test in GARCH model, the results present high degree of the autocorrelation and the ARCH effect. 6. The comparing of the models outside the sample presents the GARCH model is the better predicting model to predict the price of the futures with the stock index. Regarding predicting of the return rate of the futures with the return rate of the spots, the PDL regression model shows the better performance.