| Summary: | 碩士 === 國立高雄應用科技大學 === 商務經營研究所 === 93 === The purpose of this paper is to find out and compare with price discovery in four Taiwan spot Index, Index future, the Taiwan 50 Index and the 50 Index futures. All the data range from July 1,2004 to March 1,2005,172 are collected. The Research uses the five- minute transaction price data into four time series. The major models in this paper are Unit Root Test, Cointegration Test, Granger Causality Test, Vector Error Correction Model and Forecast Error Variance Decomposition.
Empirical results indicate that all series data are stationary with unit root test. The Research also finds that most of the spot index and futures have bi-direction Granger causality in the six groups. The results from ECM show that: In the long-term, only Taiwan 50 Index futures lead Taiwan spot index and Taiwan index future lead Taiwan 50 spot Index, others mutual leads relationship. In the short-term Taiwan index future leads Taiwan spot Index, Taiwan 50 Index and the 50 Index futures. The result shows from Forecast Error Variance Decomposition, all of variance main Forecast source is Taiwan index future. According to above all of result show that Taiwan index future is the best in price discovery.
|
|---|
