An Empirical Study of Multivariate Common Factor Model with Long-Range Dependence in Taiwan Stock Returns

• Main Authors: Sung-Nien Wu, 吳松年
• Other Authors: Nan-Jung Hsu
• Format: Others
• Language: zh-TW
• Published: 1999
• Online Access: http://ndltd.ncl.edu.tw/handle/18551946495306241003

碩士 === 東海大學 === 統計學系 === 87 === In this paper, we investigate the long-term relationship between Taiwan stock returns volatilities among different industry. The canonical correlation analysis method proposed by Ray and Tsay (1998) was used to determine the number of the common long-range dependent components. Furthermore, a multivariate common factor stochastic volatility model was used to describe the co-movement behavior between the returns volatilities among different stock indices. We used a Bayesian approach to estimate the model parameters and to predict those common long-range dependent components. We found that there are two common long-range dependence components existing among the following four industries－Foods, the Electrical industry, the Construction and the Finance. Moreover, the first common long-range dependence component mainly contributes the movement of the returns volatilities in the Electrical industry; the second common long-range dependence component mainly contributes the movement of the returns volatilities in the other three industries.