The Asymmetry and Periodieity (Holidy and Weekend Effect) of Taiwan Stock Return Rate Volatility - Application of Asymmetry Periodie - Garch Model

• Main Authors: Kuo, Fu-Ching, 郭福欽
• Other Authors: Tasi, Li-Ju
• Format: Others
• Language: zh-TW
• Published: 1997
• Online Access: http://ndltd.ncl.edu.tw/handle/98492159093346145172

碩士 === 輔仁大學 === 金融研究所 === 85 === 　　This research consists of two subjects, including (1)The asymnmetry of stock return volatility, and (2)The periodicity of stock return. The methodologies and empirical results are described as follows: 　　(1)The asymmetry of stock return volatility 　　This paper tries to find an appropriate functional form of conditional variance to describe the asymmetry of Taiwan stock volatility, by applying Hentschel''s (1995) GARCH model. It''s quite different form most of the empirical studies that arbitrarily treat some of the parameters in conditional variance equation as known a priori. Furthermore, under the free parameterization of GARCH model, we investigate whether there is still a significant relationship between return volatility and news impact of Taiwan stock. Our empirical results suggest that negtive shocks increase volatility more than positive shocks, only when shocks are very large. That is, the leverage effect is pronounced for large shocks, but not for small shocks. 　　(2)The periodicity of stock return 　　Most high-frequency asset returns exhibit periodicity volatility patterns. This paper tries to describe the periodicity of Taiwan stock return, by applying Bollerslev & Ghysels(1996) Periodic GARCH model. This new class of models feature periodicity and heteroscedasticity in the second-order moments. In order to catch periodicity and asymmetry, this paper tries to correct Bollerslev & Ghysels (1996) Periodic GARCH model by applying the asymmetry volatility patterns of Hentschel''s (1995) GARCH model. THis model is an asymmetric periodic-GARCH model, we investigate whether the asymmetric and periodic effects exist in volatility of stock market returm. Our empirical results suggest that asymmetric periodic-GARCH models can catch asymmetry of stock return volatility and periodicity stock return. Because of information accumulation, periodicity of stock return volatility exists after weekend or holiday.